Warriors rumors are in full effect as the NBA trade deadline approaches, with Golden State tied to Heat guard Jimmy Butler and Bulls center Nikola Vucevic.

Warriors trade season sounds like it's on the verge of heating up

A year ago, cybersecurity marketers were worried about the effect generative AI would have on our storytelling jobs, sparking debates about whether machines could replace human creativity.

Top 5 Cyber Trends to Track in 2025

NEC Corporation (NEC; TSE: 6701) today announced hootfolio, Inc., a provider of causal analysis AI solutions that support decision making by visualizing the cause-and-effect relationships of things from data, launched business on January 1, 2025.

hootfolio, Inc., a Provider of Causal Analysis Technology, Launches Business Development: Press Releases | NEC

The Supreme Court on Thursday revived a requirement that owners of millions of small businesses register with an arm of the Treasury Department charged with fighting money laundering and other financial crimes. The justices granted an emergency plea made by the Justice Department in the waning ...

Supreme Court allows small business registration rule to take effect, aimed at money laundering

The EU Digital Operational Resilience Act took effect last week, binding the region’s financial landscape under strict security regulations.

DORA forces EU banks to explore cybersecurity

GTA 6 is one of the most-anticipated games in history. We're all going to buy it, that much is certain, the only question that remains is how much we'll pay for it. The $70 standard for triple-A games came into effect a few years ago, but in 2024, some of the best-selling games were around $40.

As some analysts "hope" GTA 6 will normalize $100 games, one Baldur's Gate 3 dev points out that "prices haven't risen with inflation" | GamesRadar+

(Bloomberg) -- While mention of ties to Nvidia Corp. still sparks big jumps in shares of companies around the world, the gains can prove to be short-lived and some investors are questioning whether the halo effect is fading.Most Read from BloombergThese Homes Withstood the LA Fires.

Huang Bump Seen Fading for Nvidia-Linked Stocks Riding AI Rally

We recently compiled a list of the Why These 24 Stocks Are Skyrocketing. In this article, we are going to take a look at where SoundHound AI, Inc. (NASDAQ:SOUN) stands against the other stocks. One of the strongest stock market anomalies is the momentum effect, which is the tendency of recent ...

Here’s Why SoundHound AI, Inc. (SOUN) Is Skyrocketing

JP Morgan’s 5-day office mandate is expected to go into effect in March, however, employees remain staunchly opposed to the idea.

Does JP Morgan's 5-Day Office Mandate Mark the End to Remote Work?

effect

Table of Links Abstract and I. Introduction II. Background III. Related Works IV. Comparative Analysis V. Findings and Evaluation VI. Conclusion and References IV. COMPARATIVE ANALYSIS This section examines the impact on each property of the blockchain trilemma when implementing layer-one and layer-two scaling solutions across UTXO and account-based transaction models. Comparable metrics are up-scale factors, transaction throughput and dispute resolution. Qualitative comparison will be used to express the strengths and limitations of decentralisation and security. A. Layer-One Sharding The works of Luu et al. in Elastico presented a decentralised scalable solution to sharding and parallelisation, proposing feasible approaches to shard assignment without central coordination [18]. Although scaling increases with network growth linearly, individual shard security is sacrificed. This culminates in an 8% average shard failure probability, having negative security implications and undermining scaling through failure bottlenecks [21]. Resilience to byzantine threats decreases consistently, as shards are added, meaning infinite scaling eventually completely sacrifices system integrity. This can be partly attributed to small shard sizes [20]. Kokris-Kogias et al address bottleneck and security issues native to Elastico scaling through their work on OmniLedger. By increasing shard size byzantine fault tolerance is more effectively managed in a decentralised setting. The implementation of an atomic commit protocol directly combats double spending, an issue imperfectly solved in industry approaches previous to OmniLedger. Cross-shard atomicity ensures UTXO funds aren’t indefinitely locked as pending in-cases of malicious coordination. This was achieved through rejection and acceptance proof, with resilient commit and rollback parameters [20]. Probabilistically, OmniLedger could survive for 230 years before complete shard failure undermines the blockchain. In contract Elastico can survive 1 hour, under the same parameters [35]. Zilliqa utilises directory service shards to increase security and transaction efficiency through transaction coordination This mitigates double-spending and is a scalable solution to transaction sharding. Transaction speeds don’t scale as highly compared to RapidChain or OmniLedger, however transaction assignment is more secure[36]. B. Layer-Two Scalability Plasma follows a nested, parent/child chain architecture built on Ethereum [9]. Although Ethereum supports a high level of decentralisation, Plasma can centralise aspects of processing activities due to centric coordination on some sidechains [25]. Security challenges are found in consensus, weakening the overall byzantine fault tolerance of the network. For these reasons Plasma severely sacrifices security and decentralisation in the pursuit of scalability. Compared to Plasma ZKsync achieves far greater decentralisation and security at scale. Despite significantly lower transaction speeds and latency, ZKsync manages all three attributes of the blockchain trilemma while achieving impactful scalability. The zero-knowledge nature of this solution is privacy preserving by default, increasing overall system security. ZKsync is fully dependent on smart contract implementation to aggregate transaction processing, making it infeasible for UTXO based projects [32]. Nested-chains like Plasma can be deployed over a wider array of layer-one implementations, making them more feasible for less specialised use-cases. Contrasting ZKsync and zero-knowledge rollups in general, optimistic roll ups like Arbitrum focus on hyper-scalability, with disregard for other decentralisation. Supporting the highest scaling of any solution at both layer-one and two, Arbitrum is a prime example of the trilemma. The use of trusted centralised sequencers, undermines the decentralised aims of all permissionless blockchain projects. Security is increased through fraud proofs, although slow generation increases transaction latency. Although transactions are processed quickly, latency causes large time windows until a transaction is immutably confirmed. C. State Transaction Models The core UTXO account model used by Bitcoin was designed to handle transactions from one user to anothers private key address. The way the unspent transactions are generated and spent means that these transactions have a greater chance of being scaled by processing them in parallel. Although the UTXO model has minimal expressiveness in terms of on-chain logic, the UTXO model enables scalability and confidentiality, with much simpler transaction verification [11]. Vitalik introduced the Account-model into the Etheruem blockchain, which runs on a global virtual machine which maintains a record of all transactions, users and even smart contracts. Accounts are ran by private keys or smart contracts and store user assets as account balance. Although storing account state within the global state makes things easier, it also limits scalability options at layer one and requires the introduction of layer 2 solutions such as Plasma [9]. More recently, Cardano’s IOHK developed the eUTXO model, which maintains the core aspect of the UTXO model with regards to how each transaction consumes unspent outputs in order to produce new unspent transactions that can be further spent. As Cardano intended to implement smart contract functionality, using the basic UTXO model that was designed to do little more than handle payments would be unsuitable. This is highlighted through the lack of expressive coded logic available on-chain to UTXO systems. The use of an Account model addresses the issue of limited expressiveness by developing specialised smart contract accounts, becoming far to complex [37]. To create an eUTXO, two core additional pieces of functionality are needed to build upon the base UTXO model. The model needed to ensure that smart contract state was maintained and also enforce that the same smart contract logic is used throughout the whole sequence of transactions. This eUTXO model enables native-assets and smart contracts to be used on the UTXO model for the first time [6]. Authors:
(1) Cayo Fletcher-Smith, Bournemouth University;
(2) Frazer Chard. This paper is available on arxiv under CC0 1.0 DEED license. Table of Links Abstract and I. Introduction Abstract and I. Introduction II. Background II. Background III. Related Works III. Related Works IV. Comparative Analysis IV. Comparative Analysis V. Findings and Evaluation V. Findings and Evaluation VI. Conclusion and References VI. Conclusion and References IV. COMPARATIVE ANALYSIS This section examines the impact on each property of the blockchain trilemma when implementing layer-one and layer-two scaling solutions across UTXO and account-based transaction models. Comparable metrics are up-scale factors, transaction throughput and dispute resolution. Qualitative comparison will be used to express the strengths and limitations of decentralisation and security. A. Layer-One Sharding A. Layer-One Sharding The works of Luu et al. in Elastico presented a decentralised scalable solution to sharding and parallelisation, proposing feasible approaches to shard assignment without central coordination [18]. Although scaling increases with network growth linearly, individual shard security is sacrificed. This culminates in an 8% average shard failure probability, having negative security implications and undermining scaling through failure bottlenecks [21]. Resilience to byzantine threats decreases consistently, as shards are added, meaning infinite scaling eventually completely sacrifices system integrity. This can be partly attributed to small shard sizes [20]. Kokris-Kogias et al address bottleneck and security issues native to Elastico scaling through their work on OmniLedger. By increasing shard size byzantine fault tolerance is more effectively managed in a decentralised setting. The implementation of an atomic commit protocol directly combats double spending, an issue imperfectly solved in industry approaches previous to OmniLedger. Cross-shard atomicity ensures UTXO funds aren’t indefinitely locked as pending in-cases of malicious coordination. This was achieved through rejection and acceptance proof, with resilient commit and rollback parameters [20]. Probabilistically, OmniLedger could survive for 230 years before complete shard failure undermines the blockchain. In contract Elastico can survive 1 hour, under the same parameters [35]. Zilliqa utilises directory service shards to increase security and transaction efficiency through transaction coordination This mitigates double-spending and is a scalable solution to transaction sharding. Transaction speeds don’t scale as highly compared to RapidChain or OmniLedger, however transaction assignment is more secure[36]. B. Layer-Two Scalability B. Layer-Two Scalability Plasma follows a nested, parent/child chain architecture built on Ethereum [9]. Although Ethereum supports a high level of decentralisation, Plasma can centralise aspects of processing activities due to centric coordination on some sidechains [25]. Security challenges are found in consensus, weakening the overall byzantine fault tolerance of the network. For these reasons Plasma severely sacrifices security and decentralisation in the pursuit of scalability. Compared to Plasma ZKsync achieves far greater decentralisation and security at scale. Despite significantly lower transaction speeds and latency, ZKsync manages all three attributes of the blockchain trilemma while achieving impactful scalability. The zero-knowledge nature of this solution is privacy preserving by default, increasing overall system security. ZKsync is fully dependent on smart contract implementation to aggregate transaction processing, making it infeasible for UTXO based projects [32]. Nested-chains like Plasma can be deployed over a wider array of layer-one implementations, making them more feasible for less specialised use-cases. Contrasting ZKsync and zero-knowledge rollups in general, optimistic roll ups like Arbitrum focus on hyper-scalability, with disregard for other decentralisation. Supporting the highest scaling of any solution at both layer-one and two, Arbitrum is a prime example of the trilemma. The use of trusted centralised sequencers, undermines the decentralised aims of all permissionless blockchain projects. Security is increased through fraud proofs, although slow generation increases transaction latency. Although transactions are processed quickly, latency causes large time windows until a transaction is immutably confirmed. C. State Transaction Models C. State Transaction Models The core UTXO account model used by Bitcoin was designed to handle transactions from one user to anothers private key address. The way the unspent transactions are generated and spent means that these transactions have a greater chance of being scaled by processing them in parallel. Although the UTXO model has minimal expressiveness in terms of on-chain logic, the UTXO model enables scalability and confidentiality, with much simpler transaction verification [11]. Vitalik introduced the Account-model into the Etheruem blockchain, which runs on a global virtual machine which maintains a record of all transactions, users and even smart contracts. Accounts are ran by private keys or smart contracts and store user assets as account balance. Although storing account state within the global state makes things easier, it also limits scalability options at layer one and requires the introduction of layer 2 solutions such as Plasma [9]. More recently, Cardano’s IOHK developed the eUTXO model, which maintains the core aspect of the UTXO model with regards to how each transaction consumes unspent outputs in order to produce new unspent transactions that can be further spent. As Cardano intended to implement smart contract functionality, using the basic UTXO model that was designed to do little more than handle payments would be unsuitable. This is highlighted through the lack of expressive coded logic available on-chain to UTXO systems. The use of an Account model addresses the issue of limited expressiveness by developing specialised smart contract accounts, becoming far to complex [37]. To create an eUTXO, two core additional pieces of functionality are needed to build upon the base UTXO model. The model needed to ensure that smart contract state was maintained and also enforce that the same smart contract logic is used throughout the whole sequence of transactions. This eUTXO model enables native-assets and smart contracts to be used on the UTXO model for the first time [6]. Authors: (1) Cayo Fletcher-Smith, Bournemouth University; (2) Frazer Chard. Authors: Authors: (1) Cayo Fletcher-Smith, Bournemouth University; (2) Frazer Chard. This paper is available on arxiv under CC0 1.0 DEED license. This paper is available on arxiv under CC0 1.0 DEED license. available on arxiv

Layer-1 vs. Layer-2: Scaling Solutions and Their Effects on Blockchain Security and Decentralization

Why Scenario Planning Is An Effective Strategy Tool

How We Studied the Effects of Different Exoskeleton Controllers on Gait Patterns

Table of Links Abstract and 1 Introduction Methodology
Experiments
Conclusion and References 4. CONCLUSION In this paper, we propose the framework of Deep Mutual Learning across task towers(DML), which is compatible with various backbone multi-task networks. Extensive offline experiments help to verify DML’s effectiveness on multiple real-world datasets and across various base models. Moreover, thorough ablation studies are carried out to verify and understand the value of each newly introduced module. Finally, DML achieves significant online gains and has already been deployed to the online platform. REFERENCES [1] Martín Abadi, Paul Barham, Jianmin Chen, Zhifeng Chen, Andy Davis, Jeffrey Dean, Matthieu Devin, Sanjay Ghemawat, Geoffrey Irving, Michael Isard, et al. 2016. Tensorflow: a system for large-scale machine learning.. In Osdi, Vol. 16. Savannah, GA, USA, 265–283. [2] Raman Arora, Amitabh Basu, Poorya Mianjy, and Anirbit Mukherjee. 2018. Understanding Deep Neural Networks with Rectified Linear Units. In 6th International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, April 30 - May 3, 2018, Conference Track Proceedings. OpenReview.net. https://openreview.net/forum?id=B1J_rgWRW [3] Rich Caruana. 1997. Multitask learning. Machine learning 28, 1 (1997), 41–75. [4] Paul Covington, Jay Adams, and Emre Sargin. 2016. Deep neural networks for youtube recommendations. In Proceedings of the 10th ACM conference on recommender systems. 191–198. [5] Ke Ding, Xin Dong, Yong He, Lei Cheng, Chilin Fu, Zhaoxin Huan, Hai Li, Tan Yan, Liang Zhang, Xiaolu Zhang, et al. 2021. MSSM: a multiple-level sparse sharing model for efficient multi-task learning. In Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval. 2237–2241. [6] Long Duong, Trevor Cohn, Steven Bird, and Paul Cook. 2015. Low resource dependency parsing: Cross-lingual parameter sharing in a neural network parser. In Proceedings of the 53rd annual meeting of the Association for Computational Linguistics and the 7th international joint conference on natural language processing (volume 2: short papers). 845–850. [7] Peter A Flach, José Hernández-Orallo, and Cèsar Ferri Ramirez. 2011. A coherent interpretation of AUC as a measure of aggregated classification performance. In ICML. [8] F Maxwell Harper and Joseph A Konstan. 2015. The movielens datasets: History and context. Acm transactions on interactive intelligent systems (tiis) 5, 4 (2015), 1–19. [9] Robert A Jacobs, Michael I Jordan, Steven J Nowlan, and Geoffrey E Hinton. 1991. Adaptive mixtures of local experts. Neural computation 3, 1 (1991), 79–87. [10] Diederik P. Kingma and Jimmy Ba. 2015. Adam: A Method for Stochastic Optimization. In 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings, Yoshua Bengio and Yann LeCun (Eds.). http://arxiv.org/abs/1412.6980 [11] Jiaqi Ma, Zhe Zhao, Jilin Chen, Ang Li, Lichan Hong, and Ed H Chi. 2019. Snr: Sub-network routing for flexible parameter sharing in multi-task learning. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 33. 216–223. [12] Jiaqi Ma, Zhe Zhao, Xinyang Yi, Jilin Chen, Lichan Hong, and Ed H Chi. 2018. Modeling task relationships in multi-task learning with multi-gate mixture-ofexperts. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. 1930–1939. [13] Xiao Ma, Liqin Zhao, Guan Huang, Zhi Wang, Zelin Hu, Xiaoqiang Zhu, and Kun Gai. 2018. Entire space multi-task model: An effective approach for estimating post-click conversion rate. In The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval. 1137–1140. [14] Julian McAuley, Christopher Targett, Qinfeng Shi, and Anton Van Den Hengel. 2015. Image-based recommendations on styles and substitutes. In Proceedings of the 38th international ACM SIGIR conference on research and development in information retrieval. 43–52. [15] Ishan Misra, Abhinav Shrivastava, Abhinav Gupta, and Martial Hebert. 2016. Cross-stitch networks for multi-task learning. In Proceedings of the IEEE conference on computer vision and pattern recognition. 3994–4003. [16] Sebastian Ruder, Joachim Bingel, Isabelle Augenstein, and Anders Søgaard. 2019. Latent multi-task architecture learning. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 33. 4822–4829. [17] Hongyan Tang, Junning Liu, Ming Zhao, and Xudong Gong. 2020. Progressive layered extraction (ple): A novel multi-task learning (mtl) model for personalized recommendations. In Fourteenth ACM Conference on Recommender Systems. 269– 278. [18] Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. Advances in neural information processing systems 30 (2017). [19] Yining Wang, Liwei Wang, Yuanzhi Li, Di He, Wei Chen, and Tie-Yan Liu. 2013. A theoretical analysis of NDCG ranking measures. In Proceedings of the 26th annual conference on learning theory (COLT 2013), Vol. 8. 6. [20] Yuyan Wang, Zhe Zhao, Bo Dai, Christopher Fifty, Dong Lin, Lichan Hong, Li Wei, and Ed H Chi. 2022. Can Small Heads Help? Understanding and Improving Multi-Task Generalization. In Proceedings of the ACM Web Conference 2022. 3009– 3019. [21] Wikipedia. 2023. Pearson correlation coefficient — Wikipedia, The Free Encyclopedia. http://en.wikipedia.org/w/index.php?title=Pearson%20correlation% 20coefficient&oldid=1146097966. [Online; accessed 15-April-2023]. [22] Wikipedia contributors. 2022. Evaluation measures (information retrieval) — Wikipedia, The Free Encyclopedia. https://en.wikipedia.org/w/index.php?title= Evaluation_measures_(information_retrieval)&oldid=1095286224. [Online; accessed 9-January-2023]. [23] Wikipedia contributors. 2022. Mean squared error — Wikipedia, The Free Encyclopedia. https://en.wikipedia.org/w/index.php?title=Mean_squared_error& oldid=1127519968. [Online; accessed 31-January-2023]. [24] Dongbo Xi, Zhen Chen, Peng Yan, Yinger Zhang, Yongchun Zhu, Fuzhen Zhuang, and Yu Chen. 2021. Modeling the sequential dependence among audience multistep conversions with multi-task learning in targeted display advertising. In Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining. 3745–3755. [25] Yu Zhang and Qiang Yang. 2021. A survey on multi-task learning. IEEE Transactions on Knowledge and Data Engineering (2021). [26] Zhe Zhao, Lichan Hong, Li Wei, Jilin Chen, Aniruddh Nath, Shawn Andrews, Aditee Kumthekar, Maheswaran Sathiamoorthy, Xinyang Yi, and Ed Chi. 2019. Recommending what video to watch next: a multitask ranking system. In Proceedings of the 13th ACM Conference on Recommender Systems. 43–51. Authors:
(1) Yi Ren, Tencent, Beijing, China (yiren_bj@outlook.com);
(2) Ying Du, Tencent, Beijing, China (yingdu@tencent.com);
(3) Bin Wang, Tencent, Beijing, China (hillmwang@tencent.com);
(4) Shenzheng Zhang, Tencent, Beijing, China (qjzcyzhang@tencent.com). This paper is available on arxiv under CC BY 4.0 DEED license. Table of Links Abstract and 1 Introduction Abstract and 1 Introduction Methodology Experiments Conclusion and References Methodology Methodology Experiments Experiments Conclusion and References Conclusion and References 4. CONCLUSION In this paper, we propose the framework of D eep M utual L earning across task towers( DML ), which is compatible with various backbone multi-task networks. Extensive offline experiments help to verify DML’s effectiveness on multiple real-world datasets and across various base models. Moreover, thorough ablation studies are carried out to verify and understand the value of each newly introduced module. Finally, DML achieves significant online gains and has already been deployed to the online platform. D M L DML REFERENCES [1] Martín Abadi, Paul Barham, Jianmin Chen, Zhifeng Chen, Andy Davis, Jeffrey Dean, Matthieu Devin, Sanjay Ghemawat, Geoffrey Irving, Michael Isard, et al. 2016. Tensorflow: a system for large-scale machine learning.. In Osdi, Vol. 16. Savannah, GA, USA, 265–283. [2] Raman Arora, Amitabh Basu, Poorya Mianjy, and Anirbit Mukherjee. 2018. Understanding Deep Neural Networks with Rectified Linear Units. In 6th International Conference on Learning Representations, ICLR 2018, Vancouver, BC, Canada, April 30 - May 3, 2018, Conference Track Proceedings. OpenReview.net. https://openreview.net/forum?id=B1J_rgWRW [3] Rich Caruana. 1997. Multitask learning. Machine learning 28, 1 (1997), 41–75. [4] Paul Covington, Jay Adams, and Emre Sargin. 2016. Deep neural networks for youtube recommendations. In Proceedings of the 10th ACM conference on recommender systems. 191–198. [5] Ke Ding, Xin Dong, Yong He, Lei Cheng, Chilin Fu, Zhaoxin Huan, Hai Li, Tan Yan, Liang Zhang, Xiaolu Zhang, et al. 2021. MSSM: a multiple-level sparse sharing model for efficient multi-task learning. In Proceedings of the 44th International ACM SIGIR Conference on Research and Development in Information Retrieval. 2237–2241. [6] Long Duong, Trevor Cohn, Steven Bird, and Paul Cook. 2015. Low resource dependency parsing: Cross-lingual parameter sharing in a neural network parser. In Proceedings of the 53rd annual meeting of the Association for Computational Linguistics and the 7th international joint conference on natural language processing (volume 2: short papers). 845–850. [7] Peter A Flach, José Hernández-Orallo, and Cèsar Ferri Ramirez. 2011. A coherent interpretation of AUC as a measure of aggregated classification performance. In ICML. [8] F Maxwell Harper and Joseph A Konstan. 2015. The movielens datasets: History and context. Acm transactions on interactive intelligent systems (tiis) 5, 4 (2015), 1–19. [9] Robert A Jacobs, Michael I Jordan, Steven J Nowlan, and Geoffrey E Hinton. 1991. Adaptive mixtures of local experts. Neural computation 3, 1 (1991), 79–87. [10] Diederik P. Kingma and Jimmy Ba. 2015. Adam: A Method for Stochastic Optimization. In 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings, Yoshua Bengio and Yann LeCun (Eds.). http://arxiv.org/abs/1412.6980 [11] Jiaqi Ma, Zhe Zhao, Jilin Chen, Ang Li, Lichan Hong, and Ed H Chi. 2019. Snr: Sub-network routing for flexible parameter sharing in multi-task learning. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 33. 216–223. [12] Jiaqi Ma, Zhe Zhao, Xinyang Yi, Jilin Chen, Lichan Hong, and Ed H Chi. 2018. Modeling task relationships in multi-task learning with multi-gate mixture-ofexperts. In Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining. 1930–1939. [13] Xiao Ma, Liqin Zhao, Guan Huang, Zhi Wang, Zelin Hu, Xiaoqiang Zhu, and Kun Gai. 2018. Entire space multi-task model: An effective approach for estimating post-click conversion rate. In The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval. 1137–1140. [14] Julian McAuley, Christopher Targett, Qinfeng Shi, and Anton Van Den Hengel. 2015. Image-based recommendations on styles and substitutes. In Proceedings of the 38th international ACM SIGIR conference on research and development in information retrieval. 43–52. [15] Ishan Misra, Abhinav Shrivastava, Abhinav Gupta, and Martial Hebert. 2016. Cross-stitch networks for multi-task learning. In Proceedings of the IEEE conference on computer vision and pattern recognition. 3994–4003. [16] Sebastian Ruder, Joachim Bingel, Isabelle Augenstein, and Anders Søgaard. 2019. Latent multi-task architecture learning. In Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 33. 4822–4829. [17] Hongyan Tang, Junning Liu, Ming Zhao, and Xudong Gong. 2020. Progressive layered extraction (ple): A novel multi-task learning (mtl) model for personalized recommendations. In Fourteenth ACM Conference on Recommender Systems. 269– 278. [18] Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. Advances in neural information processing systems 30 (2017). [19] Yining Wang, Liwei Wang, Yuanzhi Li, Di He, Wei Chen, and Tie-Yan Liu. 2013. A theoretical analysis of NDCG ranking measures. In Proceedings of the 26th annual conference on learning theory (COLT 2013), Vol. 8. 6. [20] Yuyan Wang, Zhe Zhao, Bo Dai, Christopher Fifty, Dong Lin, Lichan Hong, Li Wei, and Ed H Chi. 2022. Can Small Heads Help? Understanding and Improving Multi-Task Generalization. In Proceedings of the ACM Web Conference 2022. 3009– 3019. [21] Wikipedia. 2023. Pearson correlation coefficient — Wikipedia, The Free Encyclopedia. http://en.wikipedia.org/w/index.php?title=Pearson%20correlation% 20coefficient&oldid=1146097966. [Online; accessed 15-April-2023]. [22] Wikipedia contributors. 2022. Evaluation measures (information retrieval) — Wikipedia, The Free Encyclopedia. https://en.wikipedia.org/w/index.php?title= Evaluation_measures_(information_retrieval)&oldid=1095286224. [Online; accessed 9-January-2023]. [23] Wikipedia contributors. 2022. Mean squared error — Wikipedia, The Free Encyclopedia. https://en.wikipedia.org/w/index.php?title=Mean_squared_error& oldid=1127519968. [Online; accessed 31-January-2023]. [24] Dongbo Xi, Zhen Chen, Peng Yan, Yinger Zhang, Yongchun Zhu, Fuzhen Zhuang, and Yu Chen. 2021. Modeling the sequential dependence among audience multistep conversions with multi-task learning in targeted display advertising. In Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining. 3745–3755. [25] Yu Zhang and Qiang Yang. 2021. A survey on multi-task learning. IEEE Transactions on Knowledge and Data Engineering (2021). [26] Zhe Zhao, Lichan Hong, Li Wei, Jilin Chen, Aniruddh Nath, Shawn Andrews, Aditee Kumthekar, Maheswaran Sathiamoorthy, Xinyang Yi, and Ed Chi. 2019. Recommending what video to watch next: a multitask ranking system. In Proceedings of the 13th ACM Conference on Recommender Systems. 43–51. Authors: (1) Yi Ren, Tencent, Beijing, China (yiren_bj@outlook.com); (2) Ying Du, Tencent, Beijing, China (yingdu@tencent.com); (3) Bin Wang, Tencent, Beijing, China (hillmwang@tencent.com); (4) Shenzheng Zhang, Tencent, Beijing, China (qjzcyzhang@tencent.com). Authors: Authors: (1) Yi Ren, Tencent, Beijing, China (yiren_bj@outlook.com); (2) Ying Du, Tencent, Beijing, China (yingdu@tencent.com); (3) Bin Wang, Tencent, Beijing, China (hillmwang@tencent.com); (4) Shenzheng Zhang, Tencent, Beijing, China (qjzcyzhang@tencent.com). This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv available on arxiv

DML Revolutionizes Multi-Task Learning with Proven Effectiveness and Real-World Deployment

The Silk Road—a marketplace that promised freedom but became a symbol of digital lawlessness—has left a legacy that continues to influence cybersecurity and legal debates today. With the recent release of key Silk Road administrator Ross Ulbricht, we are reminded of the choices that shape the balance between justice, innovation, and the evolving battle against cybercrime. The Wizard Behind the Curtain: Ross Ulbricht’s Vision and Downfall Ross Ulbricht, AKA “Dread Pirate Roberts,” envisioned the Silk Road as a libertarian experiment—a digital free market free from government control. The marketplace quickly became a hub for illegal activities, including drug trafficking, counterfeit goods, and hacking services. The Silk Road was a notorious online black market and one of the first modern darknet markets. It operated as a hidden service on the Tor network, allowing users to browse anonymously and use cryptocurrency, primarily Bitcoin, for transactions. It was designed as a marketplace for the exchange of illegal goods and services. Goods Sold on Silk Road 1.) Drugs - This was the primary product category, including: • Cannabis • Cocaine • MDMA • LSD 2.) Prescription drugs (e.g., opioids, stimulants) • Vendors often offered detailed descriptions of quality, origin, and potency, which set Silk Road apart as a marketplace. 3.) Fake Documents and Services • Fake IDs, passports, and counterfeit currency. • Hacking services, such as malware, botnets, and phishing kits. 4.) Weapons (to a limited extent) • Initially available but later banned by Ulbricht due to ethical concerns. 5.). Other Items • Software exploits and hacking tools. • Digital goods, such as stolen credit card information and account credentials • Occasional legitimate items like books or art, though this was rare. Revenue and Profit Margins Silk Road functioned on a commission-based system, taking a cut (typically 10-15%) from every transaction. The total revenue over its lifetime, Silk Road processed approximately $1.2 billion in transactions, with Ulbricht personally earning around $80 million in commissions before the site was shut down in 2013. 2. Profit Margins for Vendors. Vendors often charged a premium for the anonymity and security provided by Silk Road. Margins varied depending on many factors such as: Drug type and availability
Risks involved in shipping illegal items
Competitive pricing among vendors - vendors could mark up drugs by 200-500% compared to street prices, especially for high-demand substances or rare products. Customer pricing reflected convenience and quality assurance, as customers could review vendors and their products. This system incentivized higher-quality transactions but also increased overall costs for consumers. Down the Witches Road Ross Ulbricht was arrested on October 1, 2013, at a public library in San Francisco. The arrest marked the culmination of a complex and far-reaching investigation into the anonymous dark web marketplace he operated. The details are as follows below: Investigation Leading to Arrest 1. Initial Clues - The investigation began when an early Silk Road promotion post was traced to an online forum where the user “altoid” had advertised the marketplace in 2011. Later, the same user posted a request for tech assistance, leaving an email address tied to Ulbricht’s real identity. 2. Tracking Bitcoin - Investigators used blockchain analysis to trace transactions linked to the Silk Road, connecting them to accounts controlled by Ulbricht. 3. Operational Mistakes - Ulbricht used his personal email and real name in some communications early in the site’s existence. He once logged into a Silk Road server from his personal Wi-Fi network, exposing his IP address. 4. Physical Surveillance - The FBI monitored Ulbricht’s movements and activities in San Francisco, confirming his identity and role as DPR. Murder-for-Hire Allegations During the investigation, evidence emerged that Ulbricht had allegedly attempted to arrange six murder-for-hire plots to protect the Silk Road and its anonymity. The first allegation was in 2013 when a user named “FriendlyChemist” threatened to expose Silk Road’s vendors and customers unless paid $500,000. Ulbricht allegedly hired a hitman to kill FriendlyChemist, paying $150,000 in Bitcoin to an undercover law enforcement agent posing as the assassin. Other allegations were implicated in arranging additional hits on individuals he suspected of betraying or stealing from the Silk Road. In 2015, Ulbricht was convicted on all counts and sentenced to life in prison without the possibility of parole. The murder-for-hire allegations, while not formally prosecuted, heavily influenced the court’s perception of his character and contributed to the severity of his sentence. Ulbricht’s supporters argue that his life sentence was excessively harsh compared to sentences for similar crimes in conjunction that the murder-for-hire evidence was not robustly tested in court. In the “Eyes of Justice”, He was made an example to deter future darknet operators. The case remains one of the most high-profile examples of law enforcement’s efforts to combat cybercrime and raises ongoing debates about sentencing, privacy, and the dark web. A warning to all would-be dark web operators of their actions and potential repercussions. “Ye pirates…be warned”…. The Release: A Questionable Decision The decision to release a former Silk Road administrator into a world that has dramatically evolved since their imprisonment has raised concerns among the cybersecurity community and lawmakers alike. Much like the “Wizard” in The Wizard of Oz, the promise of the Silk Road was an illusion—one that hid the true consequences of unchecked criminal enterprise. The release of this administrator has sparked questions about accountability and whether the justice system is prepared for the potential aftershocks and left cybersecurity professionals questioning if crime truly pays. The release of a Silk Road figure may embolden other cybercriminals who see this as a sign that justice can be navigated—or even avoided. The End of the Yellow Brick Road? For the cybersecurity community, it’s a reminder that the battle against cybercrime is far from over. For lawmakers, it’s a prompt to reflect on the decisions that created a martyr out of Ross Ulbricht and a cautionary tale about the limits of justice in a digital world. The Yellow Brick Road may have led to Oz, but this story reminds us that the real wizard—the promise of freedom through technology—was always an illusion, and the choices made along the way shape the world we must now navigate. The Silk Road—a marketplace that promised freedom but became a symbol of digital lawlessness—has left a legacy that continues to influence cybersecurity and legal debates today. With the recent release of key Silk Road administrator Ross Ulbricht, we are reminded of the choices that shape the balance between justice, innovation, and the evolving battle against cybercrime. The Wizard Behind the Curtain: Ross Ulbricht’s Vision and Downfall Ross Ulbricht, AKA “Dread Pirate Roberts,” envisioned the Silk Road as a libertarian experiment—a digital free market free from government control. The marketplace quickly became a hub for illegal activities, including drug trafficking, counterfeit goods, and hacking services. The Silk Road was a notorious online black market and one of the first modern darknet markets. It operated as a hidden service on the Tor network, allowing users to browse anonymously and use cryptocurrency, primarily Bitcoin, for transactions. It was designed as a marketplace for the exchange of illegal goods and services. Goods Sold on Silk Road 1.) Drugs - This was the primary product category, including: 1.) Drugs - This was the primary product category, including: • Cannabis • Cocaine • MDMA • LSD 2.) Prescription drugs (e.g., opioids, stimulants) 2.) Prescription drugs (e.g., opioids, stimulants) • Vendors often offered detailed descriptions of quality, origin, and potency, which set Silk Road apart as a marketplace. 3.) Fake Documents and Services 3.) Fake Documents and Services • Fake IDs, passports, and counterfeit currency. • Hacking services, such as malware, botnets, and phishing kits. 4.) Weapons (to a limited extent) 4.) Weapons (to a limited extent) • Initially available but later banned by Ulbricht due to ethical concerns. 5.). Other Items • Software exploits and hacking tools. 5.). Other Items • Software exploits and hacking tools. • Digital goods, such as stolen credit card information and account credentials • Occasional legitimate items like books or art, though this was rare. Revenue and Profit Margins Revenue and Profit Margins Silk Road functioned on a commission-based system, taking a cut (typically 10-15%) from every transaction. The total revenue over its lifetime, Silk Road processed approximately $1.2 billion in transactions, with Ulbricht personally earning around $80 million in commissions before the site was shut down in 2013. 2. Profit Margins for Vendors. Vendors often charged a premium for the anonymity and security provided by Silk Road. Margins varied depending on many factors such as: Drug type and availability Risks involved in shipping illegal items Competitive pricing among vendors - vendors could mark up drugs by 200-500% compared to street prices, especially for high-demand substances or rare products. Customer pricing reflected convenience and quality assurance, as customers could review vendors and their products. This system incentivized higher-quality transactions but also increased overall costs for consumers. Drug type and availability Risks involved in shipping illegal items Competitive pricing among vendors - vendors could mark up drugs by 200-500% compared to street prices, especially for high-demand substances or rare products. Customer pricing reflected convenience and quality assurance, as customers could review vendors and their products. This system incentivized higher-quality transactions but also increased overall costs for consumers. Down the Witches Road Ross Ulbricht was arrested on October 1, 2013, at a public library in San Francisco. The arrest marked the culmination of a complex and far-reaching investigation into the anonymous dark web marketplace he operated. The details are as follows below: Investigation Leading to Arrest Investigation Leading to Arrest 1. Initial Clues - The investigation began when an early Silk Road promotion post was traced to an online forum where the user “altoid” had advertised the marketplace in 2011. Later, the same user posted a request for tech assistance, leaving an email address tied to Ulbricht’s real identity. 2. Tracking Bitcoin - Investigators used blockchain analysis to trace transactions linked to the Silk Road, connecting them to accounts controlled by Ulbricht. 3. Operational Mistakes - Ulbricht used his personal email and real name in some communications early in the site’s existence. He once logged into a Silk Road server from his personal Wi-Fi network, exposing his IP address. 4. Physical Surveillance - The FBI monitored Ulbricht’s movements and activities in San Francisco, confirming his identity and role as DPR. Murder-for-Hire Allegations Murder-for-Hire Allegations During the investigation, evidence emerged that Ulbricht had allegedly attempted to arrange six murder-for-hire plots to protect the Silk Road and its anonymity. The first allegation was in 2013 when a user named “FriendlyChemist” threatened to expose Silk Road’s vendors and customers unless paid $500,000. Ulbricht allegedly hired a hitman to kill FriendlyChemist, paying $150,000 in Bitcoin to an undercover law enforcement agent posing as the assassin. Other allegations were implicated in arranging additional hits on individuals he suspected of betraying or stealing from the Silk Road. In 2015, Ulbricht was convicted on all counts and sentenced to life in prison without the possibility of parole. The murder-for-hire allegations, while not formally prosecuted, heavily influenced the court’s perception of his character and contributed to the severity of his sentence. Ulbricht’s supporters argue that his life sentence was excessively harsh compared to sentences for similar crimes in conjunction that the murder-for-hire evidence was not robustly tested in court. In the “Eyes of Justice”, He was made an example to deter future darknet operators. The case remains one of the most high-profile examples of law enforcement’s efforts to combat cybercrime and raises ongoing debates about sentencing, privacy, and the dark web. A warning to all would-be dark web operators of their actions and potential repercussions. “Ye pirates…be warned”…. The Release: A Questionable Decision The decision to release a former Silk Road administrator into a world that has dramatically evolved since their imprisonment has raised concerns among the cybersecurity community and lawmakers alike. Much like the “Wizard” in The Wizard of Oz, the promise of the Silk Road was an illusion—one that hid the true consequences of unchecked criminal enterprise. The release of this administrator has sparked questions about accountability and whether the justice system is prepared for the potential aftershocks and left cybersecurity professionals questioning if crime truly pays. The release of a Silk Road figure may embolden other cybercriminals who see this as a sign that justice can be navigated—or even avoided. The End of the Yellow Brick Road? For the cybersecurity community, it’s a reminder that the battle against cybercrime is far from over. For lawmakers, it’s a prompt to reflect on the decisions that created a martyr out of Ross Ulbricht and a cautionary tale about the limits of justice in a digital world. The Yellow Brick Road may have led to Oz, but this story reminds us that the real wizard—the promise of freedom through technology—was always an illusion, and the choices made along the way shape the world we must now navigate.

Goodbye Yellow Brick Road: The Release of Silk Road Admin and Its Ripple Effects

Author:
(1) Prakhyat Khati, Computer Science, University of Saskatchewa, Saskatoon, Canada (khati.prakhyat@usask.com). Table of Links Abstract and 1 Introduction 2 Related Works 3 Datasets and Experiment Setup 4 Methodology 4.1 NFTs Transaction Network 4.2 NFTs Bubble Prediction 5 Discussion and Conclusions, and References 4.2 NFTs Bubble Prediction In recent years, the LPPL model[38] has been studied in different research for bubble detection, here, the model uses time-series price data. This model assumes the presence of two types of agents in the market: a group of traders with rational expectations and the second group of so-called “noise” traders, that is, irrational agents with herding behavior. The algorithm approximates the log price of the data at a given period t. as where A, B and C are the linear parameters and other are the nonlinear parameters Tc, m, w, φ. In the above equation, Tc denotes the singularity at which the previous bubble ends and transitions to another set. A denotes the log price, B denotes the magnitude of the power law acceleration and ranges from B>0 for the increasing price to B<0 for decreasing price. The LPPL model derives these positive and negative bubble indicators to all t2, thereby visualizing its own predictions. Similarly, C denotes the magnitude of the logperiodic oscillations, where if m=1, C=0, then LPPL reduces to an exponential fit of the price time series. If m<1, then LPPL grows super-exponentially until the critical time T. If C, w>0, then LppL exhibits oscillation that becomes progressively more frequent as x approaches the critical time. Here for the model, we need to calibrate the model. Model calibration can be defined as finding a unique set of model parameters that provide a good description of the system behavior and can be achieved by confronting model predictions with actual measurements performed on the system. For our test the model is calibrated with the Ordinary Least square method, providing the estimations of all parameters mention in the above equation. Similarly, the equation also contains non-linear parameters. we set the following calibration for a given time [t1, t2]. We then make bubble predictions at each period of the data, by letting this calibration iterate for the shrinking time window [t1, t2]. According to the daily data, [t1, t2] is in daily units. t2 denotes a fictitious today corresponding to t; t1 denotes an earlier day. For a given t2, the iterative calibration sets the initial range of the time window as 120 days and the shrinking Interval of t1 as 5 days. That is, we need to estimate parameters 24 times for each t2 (e.g., [1, 120], [5, 120], …, [115, 120]). This process emphasizes the prediction and here we use only historical data as input. The outcome of t2 depends only on data from t2 to the last 120 days. This paper is available on arxiv under CC BY 4.0 DEED license. Author: (1) Prakhyat Khati, Computer Science, University of Saskatchewa, Saskatoon, Canada (khati.prakhyat@usask.com). Author: Author: (1) Prakhyat Khati, Computer Science, University of Saskatchewa, Saskatoon, Canada (khati.prakhyat@usask.com). Table of Links Abstract and 1 Introduction Abstract and 1 Introduction 2 Related Works 2 Related Works 3 Datasets and Experiment Setup 3 Datasets and Experiment Setup 4 Methodology 4 Methodology 4.1 NFTs Transaction Network 4.1 NFTs Transaction Network 4.2 NFTs Bubble Prediction 4.2 NFTs Bubble Prediction 5 Discussion and Conclusions, and References 5 Discussion and Conclusions, and References 4.2 NFTs Bubble Prediction In recent years, the LPPL model[38] has been studied in different research for bubble detection, here, the model uses time-series price data. This model assumes the presence of two types of agents in the market: a group of traders with rational expectations and the second group of so-called “noise” traders, that is, irrational agents with herding behavior. The algorithm approximates the log price of the data at a given period t. as where A, B and C are the linear parameters and other are the nonlinear parameters Tc, m, w, φ. In the above equation, Tc denotes the singularity at which the previous bubble ends and transitions to another set. A denotes the log price, B denotes the magnitude of the power law acceleration and ranges from B>0 for the increasing price to B<0 for decreasing price. The LPPL model derives these positive and negative bubble indicators to all t2, thereby visualizing its own predictions. Similarly, C denotes the magnitude of the logperiodic oscillations, where if m=1, C=0, then LPPL reduces to an exponential fit of the price time series. If m<1, then LPPL grows super-exponentially until the critical time T. If C, w>0, then LppL exhibits oscillation that becomes progressively more frequent as x approaches the critical time. Here for the model, we need to calibrate the model. Model calibration can be defined as finding a unique set of model parameters that provide a good description of the system behavior and can be achieved by confronting model predictions with actual measurements performed on the system. For our test the model is calibrated with the Ordinary Least square method, providing the estimations of all parameters mention in the above equation. Similarly, the equation also contains non-linear parameters. we set the following calibration for a given time [t1, t2]. We then make bubble predictions at each period of the data, by letting this calibration iterate for the shrinking time window [t1, t2]. According to the daily data, [t1, t2] is in daily units. t2 denotes a fictitious today corresponding to t; t1 denotes an earlier day. For a given t2, the iterative calibration sets the initial range of the time window as 120 days and the shrinking Interval of t1 as 5 days. That is, we need to estimate parameters 24 times for each t2 (e.g., [1, 120], [5, 120], …, [115, 120]). This process emphasizes the prediction and here we use only historical data as input. The outcome of t2 depends only on data from t2 to the last 120 days. This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv

LPPL (Logarithmic Periodic Power Law) Model Effectively Predicts NFT Market Bubbles

Mastering Effect Programming in JavaScript with Generators and Deno

EIP-7251: Raising Maximum Effective Balance For Validators

Authors:
(1) Muhammad Zia Hydari, Katz Graduate School of Business, University of Pittsburgh and Corresponding author;
(2) Idris Adjerid, Pamplin College of Business;
(3) AAaron D. Striegel, Department of Computer Science and Engineering, University of Notre Dame. Table of Links Abstract and 1 Introduction 2. Background and 2.1. Leaderboards 3. Effect of Leaderboards on Healthful Physical Activity and 3.1. Competition 3.2. Social Influence 3.3. Moderating Effects of Prior Activity Levels and Leaderboard Size 4. Data and Model 4.1. Data 4.2. Model 5. Estimation and Robustness of the Main Effects of Leaderboards 5.2. Robustness Check for Leaderboard Initiation 5.3. Fitbit Compliance 5.4. Fitbit Attrition, Leaderboard De-Adoption, and Additional Robustness Checks 6. Heterogeneous Effect of Leaderboards 6.1. Heterogeneity by Prior Activity Levels 6.2. Interaction of Leaderboard Size, Rank, and Prior Activity Levels 6.3. Summary of Findings from Heterogeneous Effect Analysis 7. Conclusions and Discussion, Endnotes, and References 3. Effect of Leaderboards on Healthful Physical Activity Whether leaderboards will increase or decrease healthful physical activity is not entirely clear as the effect is unlikely to be similar for all individuals. Leaderboards can produce an effect on an individual’s physical activity primarily by altering this individual’s willingness to engage in physical activity. Specifically, we conjecture that changes in willingness to engage in physical activity occur primarily due to the introduction of competitive dynamics, increased individual accountability, and altering an individual’s reference point for their own activity levels. 3.1. Competition Social comparison theory suggests that a fundamental mechanism through which individuals assess their own ability is through comparison with others (Festinger 1954). Competitiveness is one manifestation of the social comparison process and drives individuals to increase their effort either ex ante to elevate their rank or ex post to maintain their high rank (Garcia et al. 2013). Thus, the first and most direct way that leaderboards impact physical activity is through the competitive dynamic that ranking a focal user against other users generates. The tag line on the Fitbit leaderboard (Figure 1)—a little healthy competition is a great thing—points to the motivational potential of this competitive mechanism. In addition, the enjoyment derived from physical activity may be impacted by the individual’s leaderboard adoption by converting the mundane activity of walking into the more exciting activity of competing against others. Therefore, individuals who may not gain any direct enjoyment from walking may engage in this activity because of the indirect enjoyment gained from competing on the leaderboard. However, prior work finds that impacts of competition on motivation and effort are highly heterogeneous and depend on several factors such as a participant’s desire to win, whether the competition provides a participant the opportunity or reason for improving their performance, and whether competition motivates a participant to put forth greater effort (Deci et al. 1981). Along this vein, a leaderboard may have minimal impact on performance if it does not provide sufficient competition or if the adopting individual is not particularly motivated by competition. More so, prior work has noted the possibility of competition having negative impacts on motivation and performance. For instance, Steinhage et al. (2015) argue that when competition elicits excitement, it may foster positive behavior. However, if competition elicits anxiety, it may foster negative behavior. Extrapolating this to our context, if the performance of others on the leaderboard elicits anxiety in the focal user, it would lead to negative outcomes for them. Reflecting this theoretical tension, the extant literature has found mixed results regarding competition with others who significantly surpass the individual in performance. Rogers and Feller (2016, p. 1) showed that “exposure to exemplary peer performances can undermine motivation and success by causing people to perceive that they cannot attain their peers’ high levels of performance,” and termed this phenomenon discouragement by peer excellence. However, Uetake and Yang (2019) find that an individual’s distance from the highest achiever has positive motivational effects, whereas comparison with the average individual has negative impacts. Thus, competition is likely a focal mechanism behind leaderboard effects but whether it positively impacts physical activity is uncertain ex ante. This paper is available on arxiv under CC BY 4.0 DEED license. Authors: (1) Muhammad Zia Hydari, Katz Graduate School of Business, University of Pittsburgh and Corresponding author; (2) Idris Adjerid, Pamplin College of Business; (3) AAaron D. Striegel, Department of Computer Science and Engineering, University of Notre Dame. Authors: Authors: (1) Muhammad Zia Hydari, Katz Graduate School of Business, University of Pittsburgh and Corresponding author; (2) Idris Adjerid, Pamplin College of Business; (3) AAaron D. Striegel, Department of Computer Science and Engineering, University of Notre Dame. Table of Links Abstract and 1 Introduction Abstract and 1 Introduction 2. Background and 2.1. Leaderboards 2. Background and 2.1. Leaderboards 3. Effect of Leaderboards on Healthful Physical Activity and 3.1. Competition 3. Effect of Leaderboards on Healthful Physical Activity and 3.1. Competition 3.2. Social Influence 3.2. Social Influence 3.3. Moderating Effects of Prior Activity Levels and Leaderboard Size 3.3. Moderating Effects of Prior Activity Levels and Leaderboard Size 4. Data and Model 4.1. Data 4.1. Data 4.2. Model 4.2. Model 5. Estimation and Robustness of the Main Effects of Leaderboards 5. Estimation and Robustness of the Main Effects of Leaderboards 5.2. Robustness Check for Leaderboard Initiation 5.2. Robustness Check for Leaderboard Initiation 5.3. Fitbit Compliance 5.3. Fitbit Compliance 5.4. Fitbit Attrition, Leaderboard De-Adoption, and Additional Robustness Checks 5.4. Fitbit Attrition, Leaderboard De-Adoption, and Additional Robustness Checks 6. Heterogeneous Effect of Leaderboards 6. Heterogeneous Effect of Leaderboards 6.1. Heterogeneity by Prior Activity Levels 6.1. Heterogeneity by Prior Activity Levels 6.2. Interaction of Leaderboard Size, Rank, and Prior Activity Levels 6.2. Interaction of Leaderboard Size, Rank, and Prior Activity Levels 6.3. Summary of Findings from Heterogeneous Effect Analysis 6.3. Summary of Findings from Heterogeneous Effect Analysis 7. Conclusions and Discussion, Endnotes, and References 7. Conclusions and Discussion, Endnotes, and References 3. Effect of Leaderboards on Healthful Physical Activity Whether leaderboards will increase or decrease healthful physical activity is not entirely clear as the effect is unlikely to be similar for all individuals. Leaderboards can produce an effect on an individual’s physical activity primarily by altering this individual’s willingness to engage in physical activity. Specifically, we conjecture that changes in willingness to engage in physical activity occur primarily due to the introduction of competitive dynamics, increased individual accountability, and altering an individual’s reference point for their own activity levels. 3.1. Competition Social comparison theory suggests that a fundamental mechanism through which individuals assess their own ability is through comparison with others (Festinger 1954). Competitiveness is one manifestation of the social comparison process and drives individuals to increase their effort either ex ante to elevate their rank or ex post to maintain their high rank (Garcia et al. 2013). Thus, the first and most direct way that leaderboards impact physical activity is through the competitive dynamic that ranking a focal user against other users generates. The tag line on the Fitbit leaderboard (Figure 1)—a little healthy competition is a great thing—points to the motivational potential of this competitive mechanism. In addition, the enjoyment derived from physical activity may be impacted by the individual’s leaderboard adoption by converting the mundane activity of walking into the more exciting activity of competing against others. Therefore, individuals who may not gain any direct enjoyment from walking may engage in this activity because of the indirect enjoyment gained from competing on the leaderboard. However, prior work finds that impacts of competition on motivation and effort are highly heterogeneous and depend on several factors such as a participant’s desire to win, whether the competition provides a participant the opportunity or reason for improving their performance, and whether competition motivates a participant to put forth greater effort (Deci et al. 1981). Along this vein, a leaderboard may have minimal impact on performance if it does not provide sufficient competition or if the adopting individual is not particularly motivated by competition. More so, prior work has noted the possibility of competition having negative impacts on motivation and performance. For instance, Steinhage et al. (2015) argue that when competition elicits excitement, it may foster positive behavior. However, if competition elicits anxiety, it may foster negative behavior. Extrapolating this to our context, if the performance of others on the leaderboard elicits anxiety in the focal user, it would lead to negative outcomes for them. Reflecting this theoretical tension, the extant literature has found mixed results regarding competition with others who significantly surpass the individual in performance. Rogers and Feller (2016, p. 1) showed that “exposure to exemplary peer performances can undermine motivation and success by causing people to perceive that they cannot attain their peers’ high levels of performance,” and termed this phenomenon discouragement by peer excellence. However, Uetake and Yang (2019) find that an individual’s distance from the highest achiever has positive motivational effects, whereas comparison with the average individual has negative impacts. Thus, competition is likely a focal mechanism behind leaderboard effects but whether it positively impacts physical activity is uncertain ex ante. This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv

Competition and Accountability: Exploring Leaderboard Effects on Physical Activity

Authors:
(1) Clauvin Almeida, Pontifical Catholic University of Rio de Janeiro (PUC-Rio), Rio de Janeiro, Brazil;
(2) Marcos Kalinowski, Pontifical Catholic University of Rio de Janeiro (PUC-Rio), Rio de Janeiro, Brazil;
(3) Anderson Uchoa, Federal University of Ceara (UFC), Itapaje, Brazil;
(4) Bruno Feijo, Pontifical Catholic University of Rio de Janeiro (PUC-Rio), Rio de Janeiro, Brazil. Table of Links Abstract and 1 Introduction 2. Background and Related Work and 2.1. Gamification 2.2. Game Design Elements and 2.3. Gamification Effects 2.4. Related Work on Gamification Negative Effects 3. Systematic Mapping and 3.1. The Research Questions 3.2. Search Strategy and 3.3. Inclusion and Exclusion Criteria 3.4. Applying the Search Strategy 3.5. Data Extraction 4. Systematic Mapping Results 5. Focus Group: Developer Perception on the Negative Effects of Game Design Elements 5.1. Context and Participant Characterization 5.2. Focus Group Design 5.3. The Developers’ Perception on The Negative Effects 5.4. On the Perceived Usefulness, Ease of use and Intent of Adoption of Mapped Negative Effects 5.5. Participant Feedback 6. Limitations 7. Concluding Remarks 7.1. Future Research Directions Acknowledgements and References 6. Limitations One of the limitations of our mapping study concerns potentially missing papers. After analyzing 3458 papers (see the list in our online repository), based on our inclusion and exclusion criteria, we initially included 79 papers. While our combined search strategy allowed identifying significantly more papers than the database search strategy employed by Toda et al. [27] (we identified 34 papers published between 2012 and 2016, against 17), the sets had differences. Therefore, we manually included the papers found by Toda et al., which were missed by our first search and extended searches, ending up with a final set of 87 papers. It is noteworthy that we verified that many of the missed papers would also have been retrieved by subsequent snowballing iterations. Nevertheless, an extension applying subsequent snowballing iterations and investigating different hybrid strategies [15], would require significant additional effort, involving analyzing several thousands of papers which could characterize a different publication instead of an extension. We are confident that our final set of included papers as part of this publication allowed providing an unbiased and meaningful overview of the adverse effects related to GDEs in gamified education software. Another risk of false negatives concerns the filtering process. We screened all papers considering only titles, abstracts, and keywords, which may not contain sufficient information to decide upon inclusion. We avoided applying EC1 and EC3 during the initial screening to lower this risk, only excluding papers that we had high confidence of not investigating GDE effects (EC1) and reporting negative effects (EC3). In case of any doubt, the paper was left for full text-based assessment. Moreover, the application of the inclusion and exclusion criteria made by the first author was reviewed by the second and fourth authors in meetings. In case of doubt, during the initial screening or full-text-based assessment, discussions were held to reach a consensus. Furthermore, we chose not to consider grey literature as part of our inclusion criteria. There is the possibility of relevant grey literature that does not have equivalent in non-grey academic papers. On the other hand, even though we did not explicitly evaluate the strength of evidence, the results herein reported are based on peer-reviewed research and backed by empirical studies. Moreover, while research that reaches negative results is essential because it shows us what does not work [115], there still seems to be a publication bias towards positive outcomes. Research reporting negative results tends to have less scientific interest, fewer citations, and to be less often published [116]. Hence, there may be additional negative results that were not published and which, for that reason, could not be included in our mapping study. Finally, in our summary of the negative effects of GDEs (cf. Figure 4), we included negative effects that may affect different roles (e.g., the students and the maintainers/teachers responsible for configuring the system and keeping it up and running with appropriate educational resources), as we found them reported in the literature. Our goal was to organize evidence regarding the overall adverse effects of GDEs in gamified education software. While we included an overview of the negative effects for different roles (e.g., RQ1.B and RQ1.C), we considered a fine-grained analysis of the effects of each GDEs for each of these roles out of the scope of our intended overview and part of future work. With regard to the focus group session, we wanted to gather insights on the mapped negative effects of game design elements from the point of view of practitioners. In a virtual focus group, generally, the number of participants is reduced, four being considered appropriate (our case) [117]. However, it is suggested to plan more online focus groups with fewer participants than when conducting face-to-face groups [118]. While we carefully planned our focus group, followed best practices, and carefully conducted our qualitative analyses, a single session is surely a limitation. Therefore, we interpret the focus group results as preliminary complementary discussions with practitioners, avoiding validity claims. Nevertheless, we had rich discussions that nicely complemented our literature study findings. It is noteworthy that these discussions already revealed interesting facts, such as the potential unawareness of developers about the negative effects of GDEs, and some example arguments on the perceived pros and cons of including the GDEs in gamified software. Still, we are aware that a single focus group is not be enough to reach generalizable findings. Unfortunately, identifying teams with experience in developing gamified software and willing to collaborate with academia is not a trivial task. Therefore, conducting new focus group sessions is part of future work. This paper is available on arxiv under CC BY 4.0 DEED license. Authors: (1) Clauvin Almeida, Pontifical Catholic University of Rio de Janeiro (PUC-Rio), Rio de Janeiro, Brazil; (2) Marcos Kalinowski, Pontifical Catholic University of Rio de Janeiro (PUC-Rio), Rio de Janeiro, Brazil; (3) Anderson Uchoa, Federal University of Ceara (UFC), Itapaje, Brazil; (4) Bruno Feijo, Pontifical Catholic University of Rio de Janeiro (PUC-Rio), Rio de Janeiro, Brazil. Authors: Authors: (1) Clauvin Almeida, Pontifical Catholic University of Rio de Janeiro (PUC-Rio), Rio de Janeiro, Brazil; (2) Marcos Kalinowski, Pontifical Catholic University of Rio de Janeiro (PUC-Rio), Rio de Janeiro, Brazil; (3) Anderson Uchoa, Federal University of Ceara (UFC), Itapaje, Brazil; (4) Bruno Feijo, Pontifical Catholic University of Rio de Janeiro (PUC-Rio), Rio de Janeiro, Brazil. Table of Links Abstract and 1 Introduction Abstract and 1 Introduction 2. Background and Related Work and 2.1. Gamification 2. Background and Related Work and 2.1. Gamification 2.2. Game Design Elements and 2.3. Gamification Effects 2.2. Game Design Elements and 2.3. Gamification Effects 2.4. Related Work on Gamification Negative Effects 2.4. Related Work on Gamification Negative Effects 3. Systematic Mapping and 3.1. The Research Questions 3. Systematic Mapping and 3.1. The Research Questions 3.2. Search Strategy and 3.3. Inclusion and Exclusion Criteria 3.2. Search Strategy and 3.3. Inclusion and Exclusion Criteria 3.4. Applying the Search Strategy 3.4. Applying the Search Strategy 3.5. Data Extraction 3.5. Data Extraction 4. Systematic Mapping Results 4. Systematic Mapping Results 5. Focus Group: Developer Perception on the Negative Effects of Game Design Elements 5. Focus Group: Developer Perception on the Negative Effects of Game Design Elements 5.1. Context and Participant Characterization 5.1. Context and Participant Characterization 5.2. Focus Group Design 5.2. Focus Group Design 5.3. The Developers’ Perception on The Negative Effects 5.3. The Developers’ Perception on The Negative Effects 5.4. On the Perceived Usefulness, Ease of use and Intent of Adoption of Mapped Negative Effects 5.4. On the Perceived Usefulness, Ease of use and Intent of Adoption of Mapped Negative Effects 5.5. Participant Feedback 5.5. Participant Feedback 6. Limitations 6. Limitations 7. Concluding Remarks 7. Concluding Remarks 7.1. Future Research Directions 7.1. Future Research Directions Acknowledgements and References Acknowledgements and References 6. Limitations One of the limitations of our mapping study concerns potentially missing papers. After analyzing 3458 papers (see the list in our online repository), based on our inclusion and exclusion criteria, we initially included 79 papers. While our combined search strategy allowed identifying significantly more papers than the database search strategy employed by Toda et al. [27] (we identified 34 papers published between 2012 and 2016, against 17), the sets had differences. Therefore, we manually included the papers found by Toda et al., which were missed by our first search and extended searches, ending up with a final set of 87 papers. It is noteworthy that we verified that many of the missed papers would also have been retrieved by subsequent snowballing iterations. Nevertheless, an extension applying subsequent snowballing iterations and investigating different hybrid strategies [15], would require significant additional effort, involving analyzing several thousands of papers which could characterize a different publication instead of an extension. We are confident that our final set of included papers as part of this publication allowed providing an unbiased and meaningful overview of the adverse effects related to GDEs in gamified education software. Another risk of false negatives concerns the filtering process. We screened all papers considering only titles, abstracts, and keywords, which may not contain sufficient information to decide upon inclusion. We avoided applying EC1 and EC3 during the initial screening to lower this risk, only excluding papers that we had high confidence of not investigating GDE effects (EC1) and reporting negative effects (EC3). In case of any doubt, the paper was left for full text-based assessment. Moreover, the application of the inclusion and exclusion criteria made by the first author was reviewed by the second and fourth authors in meetings. In case of doubt, during the initial screening or full-text-based assessment, discussions were held to reach a consensus. Furthermore, we chose not to consider grey literature as part of our inclusion criteria. There is the possibility of relevant grey literature that does not have equivalent in non-grey academic papers. On the other hand, even though we did not explicitly evaluate the strength of evidence, the results herein reported are based on peer-reviewed research and backed by empirical studies. Moreover, while research that reaches negative results is essential because it shows us what does not work [115], there still seems to be a publication bias towards positive outcomes. Research reporting negative results tends to have less scientific interest, fewer citations, and to be less often published [116]. Hence, there may be additional negative results that were not published and which, for that reason, could not be included in our mapping study. Finally, in our summary of the negative effects of GDEs (cf. Figure 4), we included negative effects that may affect different roles (e.g., the students and the maintainers/teachers responsible for configuring the system and keeping it up and running with appropriate educational resources), as we found them reported in the literature. Our goal was to organize evidence regarding the overall adverse effects of GDEs in gamified education software. While we included an overview of the negative effects for different roles (e.g., RQ1.B and RQ1.C), we considered a fine-grained analysis of the effects of each GDEs for each of these roles out of the scope of our intended overview and part of future work. With regard to the focus group session, we wanted to gather insights on the mapped negative effects of game design elements from the point of view of practitioners. In a virtual focus group, generally, the number of participants is reduced, four being considered appropriate (our case) [117]. However, it is suggested to plan more online focus groups with fewer participants than when conducting face-to-face groups [118]. While we carefully planned our focus group, followed best practices, and carefully conducted our qualitative analyses, a single session is surely a limitation. Therefore, we interpret the focus group results as preliminary complementary discussions with practitioners, avoiding validity claims. Nevertheless, we had rich discussions that nicely complemented our literature study findings. It is noteworthy that these discussions already revealed interesting facts, such as the potential unawareness of developers about the negative effects of GDEs, and some example arguments on the perceived pros and cons of including the GDEs in gamified software. Still, we are aware that a single focus group is not be enough to reach generalizable findings. Unfortunately, identifying teams with experience in developing gamified software and willing to collaborate with academia is not a trivial task. Therefore, conducting new focus group sessions is part of future work. This paper is available on arxiv under CC BY 4.0 DEED license. This paper is available on arxiv under CC BY 4.0 DEED license. available on arxiv

Stories, News & Mentions about EFFECT

Effect’s Stories on HackerNoon

Effect's News & Mentions