San Francisco is home to a melting pot of top-rated restaurants, with foods and dishes inspired by many multicultural influences. There is no denying that the food alone is reason enough to visit this beautiful city.

The 20 Best Sandwich Shops In San Francisco

It’s been a great year for local dancers who strive to keep alive the mountain tradition of clogging. The young — and young at heart — have all made their hometown proud in various competitions, and none any more so than four women representing High Country Senior Games,

Forever Young Cloggers Bring Home the Gold from State Senior Games

This popularity of what some might consider an ill-conceived and almost certainly unappealing exhibit brings up the question, Why, when most lovers of art think collectors and museum-goers like "beautiful" art,

The Allure of "Ugly" Art

NASHVILLE, Tenn. (WSMV) - One Nashville woman had hoped to start a family with her husband earlier this year. But those hopes have been put on hold for Greta Hollar after she found a lump in her breast during a self-exam in May. ”A couple weeks later, I was diagnosed with Stage 1 breast cancer,” Hollar said.

‘It will definitely be worth it all’: Nashville woman faces tough decision after breast cancer diagnosis

A former Boise middle school principal who’s been accused of failing to investigate allegations of sex abuse of a female student will face a jury trial after he pleaded not guilty to a felony charge Wednesday.

Boise principal accused of ignoring sexual abuse pleads not guilty, requests trial

Boston is known for its deep history, famous seafood, and raving sports fans — but more people should be talking about the wide variety of delicious bakeries that are scattered across this city. If you're thinking that a classic Boston cream pie is all that's on offer in this food-focused haven,

16 Best Bakeries In Boston

The trend of the real estate market in Allen County is shown by the following report, for the three weeks running Oct. 5 through 11. Figures reflect the selling price of the property, along with the address or lot number of same and are courtesy of Allen County Recorder Mona Losh’s Office.

Land transfers, Oct. 5-11

Megan Hollar scored 2 goals as the undefeated WL-S girls soccer team beat Greeneview, 3-2, on Thursday to claim its second straight OHC championship.

Tigers claim OHC girls soccer title

Hollar

hollar

CATALOGUE OF THE WORKS OF LEONARDO DA VINCI.

Will you be content if you hire and pay four people but only three show up to actually do the work? And yet, many employers do just that. According to McKinsey, an average worker spends 1.8 hours daily gathering and aggregating data, a task that is redundant and doesn’t have a direct impact on business success, and, above all, can be easily automated. Is your insurance company looking for ways to relieve your employees from this routine burden while cutting costs and minimizing errors? If so, you can consult a to build or customize an RPA solution specific to your needs. robotic process automation company But before that, where can you use RPA in insurance? And how to prepare for its implementation — either in specific processes or company wide? What is RPA, and how can it serve the insurance sector? In the insurance industry, robotic process automation (RPA) refers to rule-based software bots that can perform repetitive manual tasks, such as extracting claims data, aggregating client information from various internal and external sources, and performing background checks. If we look at the statistics of RPA implementation, the insurance sector, along with retail, are the main beneficiaries of this technology. With this popularity, Juniper Research predicts that almost a half of the insurance sector . And McKinsey believes that those who have already adopted the technology will be able to by 2025. will invest in RPA by 2024 automate 25% of their manual processes Deploying RPA in the insurance sector brings about the following benefits: . RPA can serve as a bridge between modern software, such as an ERP solution, and the existing legacy programs and devices that are still important for business. Sometimes deploying RPA in insurance can replace legacy systems altogether. For instance, a North American insurer Security Benefit in one year by automating legacy processes. Facilitates integration with legacy systems saved 40,000 labor hours . There are many real-life examples of when adopting RPA in insurance took the excessive load off employees’ backs and gave them more time for higher-value tasks. For instance, Encova Insurance saved its workers by automating their customer retention process, while Australian Unity of manual labor over the course of eight months thanks to RPA. Reduces manual work 25 hours a week spared its employees 22,493 hours . When employees are not overwhelmed with tedious routine tasks, their job satisfaction increases. A large insurance group based in Hong Kong used to delegate routine clerical work to junior staff, who made mistakes resulting in high employee turnover. The company after automating 80% of these processes. Lowers attrition rates reported lower attrition rates . Deploying RPA in insurance does require initial investment, but it helps companies to save money in the long run. For instance, a global insurance provider after deploying just 13 RPA robots. Cuts costs saved £140,000 in six months Top 7 robotic process automation insurance use cases Underwriting Underwriters must analyze information from numerous sources to assess risks and find the best rates and policy options for their clients. In the life insurance niche, for instance, this task can take up to four weeks. Underwriters of their time on data gathering and entry. spend on average 40% Bots can aggregate and process data from internal and external sites and display it on a dashboard for faster and more convenient decision making. According to McKinsey, using RPA in insurance can . Moreover, enhanced with AI, bots can perform the following tasks: reduce data processing time by 34% Populate corresponding fields in the internal system Analyze clients’ claim history and suggest pricing Verify if the person already has an existing policy Assess loss runs Identify health risks, such as checking if the person smokes Extract the client’s credit rating from third-party resources, such as Experian When RPA takes over the trivial tasks, human underwriters can focus their efforts to streamline more complex cases. A Tier 1 US insurer from different policy systems and provide underwriters with a 360-degree of the client and their risks. Prior to this collaboration, underwriters wasted a lot of time to perform this task manually. And now they could use this time to better serve the customer. turned to Accenture to automate data extraction Business process improvement The insurance sector is full of paper-intensive processes, which makes it challenging to assess operational efficiency and identify areas for improvement. When automation bots take over, it becomes possible to track their workflow and record each step. Afterwards, during audits, companies can review the logs and measure parameters, such as processing speed and the number of manual interventions required, to identify candidates for process optimization. Even by simply automating a process, the efficiency increases in terms of speed and error rate. To achieve this, many organizations are incorporating insurance RPA into their workflow. Cattolica Assicurazioni, an Italian insurance firm, turned to UiPath to automate their processes. They investigated the process of financial reconciliation and found that it contained very few exceptions and could be almost fully digitized. They deployed RPA bots to perform tedious tasks, such as matching 20.000 lines of numbers. The financial department allocated six months to complete this project. With RPA automating 90% of the process, the company with a zero error rate. accomplished this task in merely two months Claims management and fraud detection McKinsey reports that automation . reduce the amount of manual work by 80% As you can see from the stats, this is a major robotic process automation insurance use case. When claim registration is done manually, it results in errors, slow service, and unpleasant customer experience. Incorporating RPA in insurance can streamline the whole process from First Notice of Loss (FNOL) to settlement: Integrate claims data from multiple sources, such as medical records for RPA in health insurance, photos of damaged vehicles, etc. Process scanned paper claims by classifying them and entering into the system correctly Assist in claims verification and sport fraudulent claims Identify any missing information and bring this issue to the responsible officer’s attention The Canadian claims management provider SCM Insurance Services fully automated data entry for FNOL, which allowed claims . And a large US property and casualty insurer, EXL, worker compensation claim handling time by 60% within the first four months of deploying RPA in insurance claims processing. to be completed 80% faster could decrease Policy administration and cancellation Even though current software packages for policy administration save employees a lot of labor, they still include complex navigations through multiple applications, causing inefficiency and opening room for error. Insurance robotic process automation solutions, backed by and natural language processing, can receive emails from policy holders, extract data, make the required changes, such as bank mandate and address alteration, and send confirmations. The technology can handle policy rating, issuing, endorsing, renewing, and cancellation, among other tasks. machine learning Zurich Insurance Group built by Capgemini based on BluePrism software. It involves RPA robots early in policy handling as they enter policy details into the system, issue invoices, and draft policy documents that officers can review. Both Zurich clients and the insurer benefit greatly from this approach. Raffaele Nutricati, Head of Robotic Process Automation at Zurich Commercial Insurance, said, deployed an automation solution “Not only are they receiving policies of higher quality than those processed in the traditional manner, there is a significant decrease in the number of emails and phone calls to clarify information with our support desk.” Regulatory compliance The insurance sector relies on several compliance standards, such as tax law and HIPAA privacy rules, to guide documentation and processes. Officers are expected to monitor and act upon any changes in regulations. They need to align claim types with the corresponding jurisdiction as fast as possible, risking a regulatory breach. Automation bots can take over this responsibility by monitoring regulations, validating customer data accordingly, keeping a log of changes, and generating regulatory reports. For example, one popular compliance use case is Name Screening Alert Review for Sanctions on politically exposed persons (PEP). Screening systems can generate thousands of alerts every day, and it would be exhausting to verify any false positives manually. RPA in insurance can pre-process these alerts, significantly limiting the number of false positives remaining for manual verification. Customer service As we established above, deploying RPA in insurance speeds up client-facing services and almost eliminates error rates, resulting in a better customer experience. With the abundance of insurance companies, clients always have an option to leave one provider for another. According to Deloitte’s survey, after experiencing poor customer service. 41% of respondents have left their insurer Moreover, automation enables companies to perform advanced analytics to gain a better understanding of customer expectations and provide personalized offerings. For example, a client has posted on social media about his upcoming trip. An RPA assistant takes notice of this information, and when the client contacts an insurance agent about, let’s say, home insurance, the smart assistant will notify the agent about the vacation, enabling the agent to produce a customized offer covering all of the client’s needs. Such an approach will give you a competitive advantage as about the lack of personalization in insurance products. 88% of clients complain Query processing The insurance sector receives large volumes of broker and customer queries that demand fast resolution. Manual query processing can be error prone and exhausting for employees, while deploying RPA insurance bots will guarantee faster and more accurate results. Hollard Group managed to facilitate broker communication through . As a result, they reduced execution time by 600%, saving 2,000 hours of employee time per month. automating 98% of email handling processes The path to successful adoption of robotic process automation in insurance Here are the five steps that you can take to prepare yourself and your company for the upcoming automation project: . Opt for processes that are well-structured, rule-based, and don’t have many exceptions. It is also a good practice to choose frequently used processes as they will amplify the benefits and speed up return on investment. According to Gartner, the should not include more than 20 steps or involve over three applications. Identify candidate processes for automation ideal RPA candidate . Analyze the process first and eliminate any unessential steps. Also, even though every employee performs the task differently, come up with a standardized process version for automation. Optimize processes before automation . This will help people appreciate RPA benefits, give improvement suggestions, and will make them generally motivated to use the new system. Involve your employees from the beginning . There are four well-established RPA vendors, namely , , , and . Select your RPA platform UiPath Blue Prism Automation Anywhere WorkFusion . Most likely, you will still need to customize your preferred RPA solution to cover all your needs and fit into your workflows. An experienced robotic process automation company will ensure a seamless integration. Find a vendor who will help you incorporate the selected RPA insurance tools into your existing system You can learn more about in our recent article. RPA adoption challenges and how to overcome them What’s next for RPA in insurance? Deploying robotic process automation in the insurance industry can speed up operations, bring down expenses, and increase customer satisfaction. But this is just the beginning of the road towards intelligent process automation (IPA), which will empower RPA bots to take over tasks typically reserved for humans. With the help of cognitive technologies, such as , , and , RPA can maximize its contribution across the insurance value chain. AI emotion recognition optical character recognition You can learn more about intelligent automation from our recent article that . We also composed a comprehensive guide on and . Finally, you can discover more on by visiting our blog. compares IPA to RPA how to start with enterprise automation how much it will cost to implement an RPA solution RPA in finance Want to increase efficiency and cut costs by automating some of your manual insurance processes? ! Our RPA experts will help you select the right automation platform and will customize the solution to cover your unique needs. Drop us a line

RPA in Insurance: Your Complete Guide

Are you a startup? a product company? Or, a mobile app developer who wants to know about app trends? If yes, then in this post you will find this year’s top mobile app development trends that I have jotted down. After working on hundreds of projects in the past years, one thing I can say- if there is something common in every successful mobile application then it is behind it. THE ONE GREAT IDEA However, the meaning of is not the fantastic design that you have thought about or the unique feature you are planning to include in your mobile application. What I mean to say is how much value it’s going to add to your ideas. “the idea” Hence, before you start working on your next mobile application development, you should focus on quality development services and also need to ask yourself questions like - What do you need from your mobile application? What do you need to do to make it work? Who is your audience? What is the best way to develop your mobile application? One simple answer to all these questions is — . It also solves your purpose of having that one great idea for your mobile application. Keep your fingers on the pulse of the mobile app development trends Also Read: 30 Simple Mobile App Ideas for Startups (AI, ML, & AR/VR) I would like to add here is that each mobile application is different and app trends and technologies are always changing and evolving. However, what I’m sharing here is a piece of fairly standard information on the top mobile app trends to follow in 2020. I’ll go into more detail about app trends shortly. Before that, have a look at the worldwide mobile app usage stats. Global mobile application revenues are projected to generate a revenue of $188.9 billion via app stores and in-app advertising. (Refer to the below-given image) According to a region-specific revenue, in 2022, consumers in the USA are projected to spend over $34 billion dollars on mobile apps via app stores. As an expert from a software outsourcing company, I keep myself updated with the mobile app trends. Also, it’s needless to say whether you’re a mobile app developer or you run a business or looking to rent a coder, it’s important for you to stay educated in the mobile space. I’ve narrowed down the top mobile app trends for the year 2020. Here’s what you need to know. App Trend #1 - Blockchain the Game Changer Blockchain is no doubt one of the revolutionary technologies in the current scenario. But why I am calling it a game changer? Blockchain technology has helped so many startups and businesses to create their own blockchain powered apps, smart contracts, and blockchain based software solutions. In 2019, you can’t go on without blockchain. More and more techies and developer’s producing powerful apps and software using blockchain benefits. Here is the worldwide blockchain revenue estimate. Union of Blockchain IoT Blockchain in union with the Internet of Things (IoT) improves the way apps perform by speeding up transactions, reducing risk, and decreasing costs on your development project. Let’s know more about blockchain trends in 2019: Ricardian Contracts A Ricardian contract can be defined as a legal agreement that is agreed and signed upon by both of the parties involved in the contract. It can also be converted into a machine-readable contract that also defines the intentions of the involved parties. And, it’s good to use for your blockchain development related to payment and financial transactions. You can hire mobile app developers to design & develop your application on technologies like blockchain and IoT. Blockchain As A Service (BaaS) Another blockchain solution is BaaS (Blockchain As A Service), it is used by a number of business startups as well as enterprises. Basically, it allows mobile app developers with a way to link up their apps to a cloud-enabled backend space. It also provides features such as — Sending push notifications Managing users on your app Connecting to third-party cloud services Federated Blockchain Federated Blockchain is a popular blockchain trend that focuses on specific use-cases as given below: App Trend #2 — Artificial Intelligence for Your Mobile App Development Strategies Most of us from the developer’s community, have tried deep learning and machine learning for building mobile applications. Apps like , , , , Google Assistant, and IRIS aren’t new to tech developers anymore. FaceApp Replika Cortana Siri In fact, deep Learning when collaborated with machine learning can do wonders to mobile app development project by giving you valuable data and real-time analytics. Some Useful Stats on the Growth of the AI Industry - It is expected that the AI industry will rise up to $40 billion by 2020. ML industry is growing at a rate of 42% and will reach up to $9 billion by 2022. Though AI technology has become a hit launch, you still need to know about its features properly through many of its components. For instance, chatbots will play a more integral role in the deployment of AI technology. For mobile app developers, it’s now easy to integrate chatbots on their website and apps. In coming years, you’ll see chatbots more like the virtual assistants that will feel like you’re talking to a real person. Also Read: 5 Chatbot Ideas Business Should Consider in 2019 AI can also help limit toxicity in by giving the physicians with more suitable treatment options like reinforced learning that follow commands and track any form of cancer. Hire AI/ML developers to build a chatbot for your firm by integrating the features of artificial intelligence. cancer treatment App Trend #2 — VR and AR Will be Deployed Into More Mobile Apps More mobile apps will adapt to augmented reality and virtual reality in 2019. These apps incorporate the real world experience in your mobile application. The best example of an augmented reality mobile application is . Here’s the detail of the features of games- Pokemon Go All elements like grass, trees, sky, people in the game exist in the reality that you can see through your camera. But in reality, all the elements that you see through your device aren’t actually there. Important Stats According to reports, Pokemon Go has generated a revenue of $1.2 billion and was downloaded approx. 752 million times. Around 5 million active users playing Pokemon Go across the globe. It’s better to say that this was really a huge success for this gaming brand. And, that’s why more companies are going to try and use augmented reality and virtual reality apps to drive their mobile application downloads and revenue for their business. App Trend #3 — 5G Technology for Mobile App Developers 5G technology is going to be the most popular app development trend for the year 2019. For mobile app developers, technology is about to transform mobility forever. Thousands of app developers are taking into consideration to enhance the functionality of their apps in relation to data security, speed, 3D gaming, augmented reality and adding new features to their mobile application. 5G mobile technology Development related to these areas will be greatly influenced by 5G. With that said, developers should expect to see more users on their mobile app, as you will provide them with better and faster top-end speeds than ever before. The features such as custom made network slices and ultra-low network latency can help improve your mobile application responsiveness and speed that is something beyond just a slower internet connection. , For example AT&T has recently launched its AT&T 5G Innovation Program to continue to create and develop new customer experiences with the power of 5G. The program is for mobile application developers to bring 5G ideas into experiences of life. I also recommend app developers use 5G network speed when developing and testing their mobile apps. It is important because apps developed exclusively in areas with fast Wi-Fi and high-speed cellular services such as 5G operate satisfactorily for users with limited speed options. App Trend #4 — Instant Apps for Mobile App Development The concept of instant mobile apps is getting more popular day by day. The best part is, instant apps are not only popular among individuals, but businesses also finding them useful because it helps achieve a better conversion rate in comparison to a regular mobile app. iOS and Android app developers For example, apps like Hollar and NYTimes have increased their conversion rate by 20% and 27% respectively by using the instant mobile application. Additionally, top development brands of the gaming industry like and Games have also improved their player acquisition rate by using the Instant mobile application. King Hothead In basic terms, instant mobile apps are the type of native mobile application that you don’t require to download and install on your Android or iOS enabled smartphone. They are smaller in size when compared to regular mobile apps.These apps provide the basic functionalities of a website. Deliver a great user experience does not consume extra space in your phone’s memory. A few more examples of such apps include — and . And, 2019 will have more instant apps emerging for both developers and individuals. Hollar Inc. , New York Times Crossword , Red Bull TV , Wish — Shopping Made Fun , Skyscanner App Trend #5 — IoT Trends In the Sphere of Mobile App Development The Internet of Things for mobile app development has become a big industry that is evolving with time. Using this technology, mobile application developers can create intuitive platforms to control different non-IT equipment via apps. Almost all tech giants have started developing IoT friendly apps. It is expected that the IoT investment will reach $120 billion by 2020. IoT mobile app developers have to select a viable platform for app development. So what should be the best strategy? To help you with this, I have curated a few points here: — The mobile app developers can use the Android platform to create an account that can be connected to Ubidots. By doing so, you’ll be able to push your app on smartphones. Android Ubidots — This platform provides you with the IoT like features for your app development project. Xively — This platform provides you with an advanced level of tools to market your mobile application. Generally, Thingworx is popular because of the following benefits: Thingworx It’s easy to segregate products and services. Support optimization in your business activities.Improve the overall user experience on your mobile app. Help you build and identify new revenue channels. Another best platform for the mobile app developers is the Forester Wave Report that is helpful in creating viable IoT strategies for your app development project. Or it is good to hire mobile app developers to make your app stand out in the app market. App Trend #6 — 2020 the Year of Mobile Wallets In 2019, customers will shift towards m-commerce, as technologies like IoT and decentralized platforms are entering into the mobile payments and powering the mobile apps with more secured transaction options. The brand like Samsung is using mobile wallets to give their customers a more convenient and safe platform for money transaction and bill payments. In addition, encrypted apps ensure the users’ money remains safe from possible theft and vulnerabilities. Mobile wallet app trends are in demand due to internet payment systems that require fast and secured channels for money transactions. Samsung mobile wallet For example, is the best way of making digital transactions for the peer-to-peer money payment system. Apple Pay in iOS 11 Seems a new thing altogether? You can hire Android developers for executing these features. Final Words A million of the mobile apps are present in the play store. You have to face the struggle to stand out in this competitive space. To make your mobile app development successful it is important to keep yourself updated with the latest mobile app development trends and technologies. And I hope that with these points that I’ve shared here give you a complete insight of top mobile development trends to follow in 2020. Also, while it would be a bit difficult for businesses to identify the exact benefits of all these mobile application development trends, and if that’s your case, just reach out to experts of the industry.  Or you can contact top global IT solutions company to work on the creation of your next business application.

Top Mobile App Development Trends of the Year 2020

Crayon Graph, from the Google Founders Collection at the Computer History Museum Preserving a Culture of Innovation In May 1998, angel investor Ron Conway first learned about at Palo Alto’s Empire Tap Room, now closed. On the receipt, he prophetically inscribed “this is going to be huge!” Google This receipt is part of a trove of early artifacts and digital assets coming to the Computer History Museum (CHM) in what’s been dubbed the “Google Founders Collection.” A cornerstone acquisition of CHM’s Exponential Center, dedicated to capturing the legacy and advancing the entrepreneurship and innovation, the collection contains objects, documents, photographs, and video from Google’s first 10 years. future “Google embodies the spirit of entrepreneurship and innovation that the Exponential Center is documenting, interpreting, and sharing with people around the world,” said , executive director of the Exponential Center. “This collection will support the production of many rich lessons of risk-taking, innovation, and bold ideas that will both tell the early Google story and inspire future entrepreneurs.” Marguerite Gong Hancock Other notable artifacts include Google’s first business plan and the “Crayon Graph,” first drawn by Lucas Pereira to track query growth and then added to by other Googlers to highlight milestones and special events. Follow Computer History Museum | Facebook Twitter The initial collection was assembled within Google by VP of Product Management Richard Holden under the leadership of Susan Wojcicki, now CEO of YouTube, who housed Google’s early-stage operation — led by co-founders Larry Page and Sergey Brin — in her home in Menlo Park, California. (Wojcicki later served as a senior vice president of Google, before moving to YouTube.) Curatorial Director of CHM’s Internet History Program worked closely with Wojcicki and Google, serving as the project’s web historian, and provided essential historical interpretation and scope. Marc Weber “We are delighted to be collaborating with Google to protect and preserve this pivotal part of Silicon Valley history, and grateful to Susan Wojcicki for demonstrating that a culture of preservation within these historic companies can be as important to future generations as a culture of innovation.” — , CEO & President, Computer History Museum John Hollar The Museum’s Collections team will lead the work of cataloging and archiving the collection, which is expected to take a number of years. In continued collaboration with Google, the Museum will also be accepting future artifact donations that are relevant to the startup decade covered by the collection. While the Museum with the Exponential Center may develop future programs based on the collection will not currently be accessible by the public. “The Computer History Museum’s impressive collections and archiving expertise made the Museum a natural choice to preserve the Google Founders Collection. We look forward to collaborating with CHM as the collection expands.” — Susan Wojcicki, CEO of YouTube Establishing a Culture of Preservation This network engineer cap inspired the Cisco cultural tradition of unique hats as give-away items at the company’s yearly users symposium, previously Networkers — now Cisco Live! Center for Cisco Heritage, managed by the Computer History Museum. The Google Founders Collection joins a number of corporate and organization archives preserved or managed by the Museum, including early computing giant Digital Equipment Corporation and social media precursor . These collections comprise early documents, photos, records, media, and artifacts that embody the startup spirit of early Silicon Valley and a budding technology industry. They represent a culture of innovation. Community Memory Most recently, CHM and Cisco, in a groundbreaking collaboration, launched the , which emphasizes the importance of preserving Silicon Valley’s corporate history and presents a new model for open business archives. Led by CHM’s Paula Jabloner and Stephanie Waslohn, the center comprises a collection of artifacts, documents, media, and ephemera — from early campaign posters to some of the company's first products— that tell the story of Cisco’s 30-year history. The three-year project culminated with the exhibit , which opened on March 31, 2017, to the cheers of former and current employees, board members, and executives, including former Cisco CEOs John Chambers and John Morgridge. In 2016 the center became its own LLC, with Museum trustee and Cisco executive Don Proctor as its CEO, and endowed to provide long-term, sustainable preservation. Center for Cisco Heritage Our Story In the coming weeks, the Museum will also release the largest and most complete set of records from Digital Equipment Corporation (DEC), dating from 1947 through 2002, with the bulk from the company’s years of operation from 1957 through 1998, when they were bought by Compaq Computer. Founded in 1957 by Ken Olsen and Harlan Anderson, DEC was one of the most successful computer companies of the 20th century, reaching nearly $14 billion in revenue and over 120,000 employees worldwide. Community Memory terminal at Leopold’s Records, Berkeley, California, ca. 1975. Collection of the Computer History Museum, Community Memory records, Box 12, Folder 15, Catalog . 102703229

“This Is Going to Be Huge”: Google Founders Collection Comes to Computer History Museum

Startups of The Year: Meet the Media Industry

Startups des Jahres: Lernen Sie die Medienbranche kennen

Startups της χρονιάς: Γνωρίστε τη βιομηχανία των μέσων ενημέρωσης

Startups del año: conozca la industria de los medios

Urteko startup-ak: ezagutu komunikabideen industria

今年のスタートアップ: メディア業界を知る

Metų startuoliai: susipažinkite su žiniasklaidos pramone

Оны шилдэг гарааны бизнесүүд: Хэвлэл мэдээллийн салбарынхантай танилц

Startups van het jaar: maak kennis met de media-industrie

Startupy Roku: Poznaj Branżę Medialną

د کال پیل: د رسنیو صنعت سره لیدنه

වසරේ ආරම්භක: මාධ්‍ය කර්මාන්තය හමුවන්න

Startups of The Year: Hlangana na Vumaki bya Vuhangalasi bya Mahungu

Tonomy Foundation has been nominated in HackerNoon's annual Startups of The Year awards in Amsterdam, Netherlands. Please vote for us here: Region: https://hackernoon.com/startups/europe/europe-bussum-north-holland-netherlands?stup=66f327b36ea4677e23e48baf Software Development: https://hackernoon.com/startups/industry/software-development?stup=66f327b36ea4677e23e48baf SaaS: https://hackernoon.com/startups/industry/saas?stup=66f327b36ea4677e23e48baf IT Services: https://hackernoon.com/startups/industry/it-services?stup=66f327b36ea4677e23e48baf Meet Tonomy Foundation: Building Pangea Virtual Nation Tonomy Foundation is pioneering the future of decentralized governance and digital identity. Our flagship project, Pangea, is a sovereign virtual nation built on Web 4.0, enabling transparent governance, secure digital identity, and a resilient global economy powered by the LEOS (Ƚ) currency. How Pangea achieved product-market fit Tonomy achieved product-market fit by addressing the growing global demand for secure, decentralized governance and identity solutions. Our ecosystem solves critical challenges in privacy, cybersecurity, and usability that traditional Web3 infrastructures face, providing a seamless and user-friendly experience with advanced AI integration and zero-knowledge architecture. We have demonstrated strong adoption from early tech innovators, blockchain developers, and decentralized autonomous organizations (DAOs), who benefit from our scalable infrastructure, compliance capabilities, and the Pangea Passport's passwordless identity system. The evolution of Pangea Building Pangea required several iterations, from refining our identity management tools to perfecting the multi-layer governance framework. Early on, we identified gaps in Web3's scalability and security, which led us to pivot toward Web 4.0. User feedback helped us focus on creating a highly customizable governance system that balances privacy with transparency. This pivot led to the creation of Pangea, a first-of-its-kind virtual nation. Milestones and achievements Launched Pangea, the first global decentralized governance platform.
Developed LEOS, a MICA-compliant digital currency designed for seamless global payments and governance participation.
Partnered with the World Wide Web Consortium (W3C) on decentralized identity standards.
Won 2nd place at the Odyssey Hackathon for our participatory budgeting tool, Pangea Participate. Lessons learned Our biggest challenges have revolved around navigating complex regulatory landscapes and ensuring that our solutions are accessible to both technical and non-technical users. We learned that building strong partnerships, particularly in the legal and compliance space, is crucial for long-term success. Additionally, creating a robust yet flexible platform was key to accommodating diverse use cases across industries. What Startups of The Year means to us Participating in HackerNoon’s Startups of the Year provides us with the opportunity to showcase how decentralized governance and digital identity can transform the global economy. We aim to connect with like-minded innovators, expand our community, and drive the adoption of Web 4.0 technologies. Conclusion Vote for Tonomy Foundation today! Help us lead the charge in building a more secure, decentralized, and inclusive digital world. Tonomy Foundation has been nominated in HackerNoon's annual Startups of The Year awards in Amsterdam , Netherlands . Amsterdam Netherlands Please vote for us here: Please vote for us here: Region: https://hackernoon.com/startups/europe/europe-bussum-north-holland-netherlands?stup=66f327b36ea4677e23e48baf https://hackernoon.com/startups/europe/europe-bussum-north-holland-netherlands?stup=66f327b36ea4677e23e48baf Software Development: https://hackernoon.com/startups/industry/software-development?stup=66f327b36ea4677e23e48baf https://hackernoon.com/startups/industry/software-development?stup=66f327b36ea4677e23e48baf SaaS: https://hackernoon.com/startups/industry/saas?stup=66f327b36ea4677e23e48baf https://hackernoon.com/startups/industry/saas?stup=66f327b36ea4677e23e48baf IT Services: https://hackernoon.com/startups/industry/it-services?stup=66f327b36ea4677e23e48baf https://hackernoon.com/startups/industry/it-services?stup=66f327b36ea4677e23e48baf Meet Tonomy Foundation: Building Pangea Virtual Nation Meet Tonomy Foundation: Building Pangea Virtual Nation Tonomy Foundation is pioneering the future of decentralized governance and digital identity. Our flagship project, Pangea , is a sovereign virtual nation built on Web 4.0, enabling transparent governance, secure digital identity, and a resilient global economy powered by the LEOS (Ƚ) currency. Pangea Pangea How Pangea achieved product-market fit How Pangea achieved product-market fit Tonomy achieved product-market fit by addressing the growing global demand for secure, decentralized governance and identity solutions. Our ecosystem solves critical challenges in privacy, cybersecurity, and usability that traditional Web3 infrastructures face, providing a seamless and user-friendly experience with advanced AI integration and zero-knowledge architecture. We have demonstrated strong adoption from early tech innovators, blockchain developers, and decentralized autonomous organizations (DAOs), who benefit from our scalable infrastructure, compliance capabilities, and the Pangea Passport's passwordless identity system. The evolution of Pangea The evolution of Pangea Building Pangea required several iterations, from refining our identity management tools to perfecting the multi-layer governance framework. Early on, we identified gaps in Web3's scalability and security, which led us to pivot toward Web 4.0. User feedback helped us focus on creating a highly customizable governance system that balances privacy with transparency. This pivot led to the creation of Pangea, a first-of-its-kind virtual nation. Milestones and achievements Milestones and achievements Launched Pangea, the first global decentralized governance platform. Developed LEOS, a MICA-compliant digital currency designed for seamless global payments and governance participation. Partnered with the World Wide Web Consortium (W3C) on decentralized identity standards. Won 2nd place at the Odyssey Hackathon for our participatory budgeting tool, Pangea Participate. Launched Pangea , the first global decentralized governance platform. Pangea Developed LEOS , a MICA-compliant digital currency designed for seamless global payments and governance participation. LEOS Partnered with the World Wide Web Consortium (W3C) on decentralized identity standards. Won 2nd place at the Odyssey Hackathon for our participatory budgeting tool, Pangea Participate . Pangea Participate Lessons learned Lessons learned Our biggest challenges have revolved around navigating complex regulatory landscapes and ensuring that our solutions are accessible to both technical and non-technical users. We learned that building strong partnerships, particularly in the legal and compliance space, is crucial for long-term success. Additionally, creating a robust yet flexible platform was key to accommodating diverse use cases across industries. What Startups of The Year means to us What Startups of The Year means to us Participating in HackerNoon’s Startups of the Year provides us with the opportunity to showcase how decentralized governance and digital identity can transform the global economy. We aim to connect with like-minded innovators, expand our community, and drive the adoption of Web 4.0 technologies. Conclusion Conclusion Vote for Tonomy Foundation today! Help us lead the charge in building a more secure, decentralized, and inclusive digital world.

Introducing the Tonomy Foundation and Pangea Virtual Nation, Startups of the Year 2024 Nominee

I actually planned and wanted to write a story about the Joker 2.0 movie, but as Elon Musk Xed himself, some memes/stories write themselves: So, here we are. The gloves are off, or it’s better to say, a mask is off. This is not my first story about Elon Musk on HackerNoon. And, it’s not going to be the last. Elon Musk saying at the Butler Rally that this would be the ‘Last Election’ if Trump loses, is just like me writing that this is the most important story in tech history, and if it’s not the top story it’d be my last. So, let’s just chill. OK? I look back and realize that my stories about Elon were more about profiling than writing. That’s why I’m not surprised at all. Why did he do it? Who? Joaquin Phoenix or Elon Musk? Joaquin ‘Commodus’ Phoenix pissed on his Oscar by appearing in Joker 2 for the money. “He was paid a whopping $20 million upfront.” Trump’s rally in Butler turned out to be a musical just like “Joker: Folie à Deux,” but Elon X Musk didn’t do it for the money. Once you have fuck-you-money, you don’t need money. Trump shared a nice and convenient story about his rich friend and how he helped him get a sit at the table of some prestigious restaurant or something by the power of his voice before making an intro for Elon before the Xtatic crowd in Butler. Trump Blinks, Here Comes A Fleet of Starlinks No Trump. No Internet for Hurricane Helene victims. Am I right? Because He called Elon, and not the other way around. Let’s say I’m being too harsh with the whole Starlink thing. The man who has once been bitten by the snake fears every piece of rope (Chinese proverb). I still remember what I wrote about Elon and a Ukrainian government official. What I do not remember is how, when, and why Starlink in Ukraine got to a ‘now you see me, now you don't situation’. I mean, it’s a long way from “some” Ukrainian official and Kim Dotcom to Trump, isn’t it? Nah, Elon, you ain’t damned, you’re just being pragmatic as always. I can only imagine the MAGA MEGA “lengths” to which the small city of Bastrop would go to help the new X’s HQ. Thus Spoke Zarathustra, Thus Spoke Musk, Per Aspera Ad Astra I’m not an American, but I realize that Musk already has his own version of the “Plan B:” So, my dear friends and readers in America, “if Trump is NOT elected,” pack your bags because you’re going to Mars. You are damned if he’s elected, you are damned if he’s not elected, so both Elon and I guess, you’re just damned. People can lose their minds when they flip sides. I’m just sayin’. “Within the context of American politics, Musk has said he supported Barack Obama in 2008 and 2012, Hillary Clinton in 2016, Joe Biden in 2020, and Donald Trump in 2024.” “Liz Cheney warns Trump will never leave office if he's elected president again. Asked if she believed Trump would try to stay in power forever, Cheney said, Absolutely.” Hear, hear! Musk won’t decide who’s going to be the 47th President of the United States, but Trump will definitely decide who’s going to be the 48th President. Lemme give you a hint. He’s a huge LOTR fan and he read all of Asimov’s “Foundation” books. I actually planned and wanted to write a story about the Joker 2.0 movie, but as Elon Musk Xed himself, some memes/stories write themselves: So, here we are. The gloves are off, or it’s better to say, a mask is off. This is not my first story about Elon Musk on HackerNoon. And, it’s not going to be the last. Elon Musk saying at the Butler Rally that this would be the ‘Last Election’ if Trump loses , is just like me writing that this is the most important story in tech history, and if it’s not the top story it’d be my last. So, let’s just chill. OK? the ‘Last Election’ if Trump loses I look back and realize that my stories about Elon were more about profiling than writing. That’s why I’m not surprised at all. Why did he do it? Who? Joaquin Phoenix or Elon Musk? Joaquin ‘Commodus’ Phoenix pissed on his Oscar by appearing in Joker 2 for the money. “He was paid a whopping $20 million upfront .” Trump’s rally in Butler turned out to be a musical just like “Joker: Folie à Deux,” but Elon X Musk didn’t do it for the money. Once you have fuck-you-money, you don’t need money. Trump shared a nice and convenient story about his rich friend and how he helped him get a sit at the table of some prestigious restaurant or something by the power of his voice before making an intro for Elon before the Xtatic crowd in Butler. $20 million upfront $20 million upfront Trump Blinks, Here Comes A Fleet of Starlinks No Trump. No Internet for Hurricane Helene victims. Am I right? Because He called Elon, and not the other way around. Let’s say I’m being too harsh with the whole Starlink thing. The man who has once been bitten by the snake fears every piece of rope (Chinese proverb). I still remember what I wrote about Elon and a Ukrainian government official . Elon and a Ukrainian government official What I do not remember is how, when, and why Starlink in Ukraine got to a ‘now you see me, now you don't situation’. I mean, it’s a long way from “some” Ukrainian official and Kim Dotcom to Trump, isn’t it? Kim Dotcom Nah, Elon, you ain’t damned, you’re just being pragmatic as always. being pragmatic I can only imagine the MAGA MEGA “lengths” to which the small city of Bastrop would go to help the new X’s HQ . the new X’s HQ Thus Spoke Zarathustra, Thus Spoke Musk, Per Aspera Ad Astra I’m not an American, but I realize that Musk already has his own version of the “Plan B:” So, my dear friends and readers in America, “if Trump is NOT elected,” pack your bags because you’re going to Mars. You are damned if he’s elected, you are damned if he’s not elected, so both Elon and I guess, you’re just damned. People can lose their minds when they flip sides. I’m just sayin’. “Within the context of American politics, Musk has said he supported Barack Obama in 2008 and 2012, Hillary Clinton in 2016, Joe Biden in 2020, and Donald Trump in 2024.” Musk has said he supported Barack Obama “Liz Cheney warns Trump will never leave office if he's elected president again. Asked if she believed Trump would try to stay in power forever, Cheney said, Absolutely.” Trump will never leave office Hear, hear! Musk won’t decide who’s going to be the 47th President of the United States , but Trump will definitely decide who’s going to be the 48th President. Lemme give you a hint. He’s a huge LOTR fan and he read all of Asimov’s “Foundation” books. the 47th President of the United States

Elon Musk SPQR - Starlink Populus Que Republicanus 

Controlling Automated Vehicles on Large Lane-free Roundabouts (Extended Version): References

Author:
(1) Attila Losonczi. Table of Links Abstract and 1 Introduction 1.1 Basic notions and notations 1.2 Basic definitions from [7] and [8] 2 Generalized integral 2.1 Multiplication on [0, +∞) × [−∞, +∞] 2.2 Measurability 2.3 The integral of functions taking values in [0, +∞) × [0, +∞) 3 Applications 4 References 4 References [1] Billingsley, P., Probability and Measure, John Wiley & Sons, vol. 939 (2012). [2] Bruckner, A. M.; Bruckner, J. B.; Thomson, B. S., Real Analysis, CreateSpace Independent Publishing Platform; 2nd edition (July 31, 2008) [3] Engelking, R., Dimension theory, North-Holland Publishing Company (1978). [4] Falconer, Kenneth, Fractal geometry: mathematical foundations and applications, John Wiley & Sons (2004). [5] Kurtz, Douglas S.; Swartz, Charles W., Series in Real Analysis, Theories of integration: The Integrals of Riemann, Lebesgue, Henstock-Kurzweil, and McShane, World Scientific Publishing Company (2011). [6] Losonczi, Attila, Dimension structures, Journal of Classical Analysis 14, Number 1 (2019), 69–95. [7] Losonczi, Attila, The Hausdorff-integral and its applications, arXiv preprint, http://arxiv.org/abs/2401.11465. [8] Losonczi, Attila, The Hausdorff-integral on h-measure spaces, arXiv preprint, https://arxiv.org/abs/2402.09118v2. [9] Losonczi, Attila, Measuring the deficiency of properties in analysis, topology and geometry, (2023), in preparation This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. Author: (1) Attila Losonczi. Author: Author: (1) Attila Losonczi. Table of Links Abstract and 1 Introduction Abstract and 1 Introduction 1.1 Basic notions and notations 1.1 Basic notions and notations 1.2 Basic definitions from [7] and [8] 1.2 Basic definitions from [7] and [8] 2 Generalized integral 2.1 Multiplication on [0, +∞) × [−∞, +∞] 2.1 Multiplication on [0, +∞) × [−∞, +∞] 2.2 Measurability 2.2 Measurability 2.3 The integral of functions taking values in [0, +∞) × [0, +∞) 2.3 The integral of functions taking values in [0, +∞) × [0, +∞) 3 Applications 3 Applications 4 References 4 References 4 References [1] Billingsley, P., Probability and Measure, John Wiley & Sons, vol. 939 (2012). [2] Bruckner, A. M.; Bruckner, J. B.; Thomson, B. S., Real Analysis, CreateSpace Independent Publishing Platform; 2nd edition (July 31, 2008) [3] Engelking, R., Dimension theory, North-Holland Publishing Company (1978). [4] Falconer, Kenneth, Fractal geometry: mathematical foundations and applications, John Wiley & Sons (2004). [5] Kurtz, Douglas S.; Swartz, Charles W., Series in Real Analysis, Theories of integration: The Integrals of Riemann, Lebesgue, Henstock-Kurzweil, and McShane, World Scientific Publishing Company (2011). [6] Losonczi, Attila, Dimension structures, Journal of Classical Analysis 14, Number 1 (2019), 69–95. [7] Losonczi, Attila, The Hausdorff-integral and its applications, arXiv preprint, http://arxiv.org/abs/2401.11465. [8] Losonczi, Attila, The Hausdorff-integral on h-measure spaces, arXiv preprint, https://arxiv.org/abs/2402.09118v2. [9] Losonczi, Attila, Measuring the deficiency of properties in analysis, topology and geometry, (2023), in preparation This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. This paper is available on arxiv under CC BY-NC-ND 4.0 DEED license. available on arxiv

Generalized Hausdorff Integral and Its Applications: References

Authors:
(1) Troisemaine Colin, Department of Computer Science, IMT Atlantique, Brest, France., and Orange Labs, Lannion, France;
(2) Reiffers-Masson Alexandre, Department of Computer Science, IMT Atlantique, Brest, France.;
(3) Gosselin Stephane, Orange Labs, Lannion, France;
(4) Lemaire Vincent, Orange Labs, Lannion, France;
(5) Vaton Sandrine, Department of Computer Science, IMT Atlantique, Brest, France. Table of Links Abstract and Intro Related work Approaches Hyperparameter optimization Estimating the number of novel classes Full training procedure Experiments Conclusion Declarations References Appendix A: Additional result metrics Appendix B: Hyperparameters Appendix C: Cluster Validity Indices numerical results Appendix D: NCD k-means centroids convergence study 8 Conclusion In this article, we have shown that in the NCD setting, unsupervised clustering algorithms can benefit from knowledge of the known classes and reliably improve their performance by implementing simple modifications. We have also introduced a novel NCD algorithm called PBN, which is characterized by its simplicity and low number of hyperparameters, which proved to be a decisive advantage under realistic conditions. In addition, we have proposed an adaptation of the k-fold cross-validation process to tune the hyperparameters of NCD methods without depending on the labels of the novel classes. Finally, we have demonstrated that the number of novel classes can be accurately estimated within the latent space of PBN. These two previous contributions have shown that the NCD problem can be solved in realistic situations where no prior knowledge of the novel classes is available during training. Declarations Funding Colin Troisemaine, Alexandre Reiffers-Masson, Stephane Gosselin, Vincent Lemaire and Sandrine Vaton received funding from Orange SA. Competing Interests Colin Troisemaine, St´ephane Gosselin and Vincent Lemaire received research support from Orange SA. Alexandre Reiffers-Masson and Sandrine Vaton received research support from IMT Atlantique. Ethics approval Not applicable. Consent to participate All authors have read and approved the final manuscript. Consent for publication Not applicable. Availability of data and materials All data used is this study are available publicly online. The datasets were extracted directly in the repositories available with the links in the corresponding section. Code availability The code for experiments is available at the following url: https://github.com/ PracticalNCD/ECMLPKDD2024. Authors’ contributions Colin Troisemaine, Alexandre Reiffers-Masson, St´ephane Gosselin, Vincent Lemaire and Sandrine Vaton contributed to the manuscript equally. References [1] Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., Fei-Fei, L.: Imagenet: A largescale hierarchical image database. In: CVPR, pp. 248–255 (2009) [2] Troisemaine, C., Lemaire, V., Gosselin, S., Reiffers-Masson, A., Flocon-Cholet, J., Vaton, S.: Novel class discovery: an introduction and key concepts. ArXiv (2023) [3] Hsu, Y.-C., Lv, Z., Kira, Z.: Learning to cluster in order to transfer across domains and tasks. In: ICLR (2018) [4] Han, K., Rebuffi, S.-A., Ehrhardt, S., Vedaldi, A., Zisserman, A.: Autonovel: Automatically discovering and learning novel visual categories. PAMI (2021) [5] Zhong, Z., Zhu, L., Luo, Z., Li, S., Yang, Y., Sebe, N.: Openmix: Reviving known knowledge for discovering novel visual categories in an open world. In: CVPR, pp. 9462–9470 (2021) [6] Zhong, Z., Fini, E., Roy, S., Luo, Z., Ricci, E., Sebe, N.: Neighborhood contrastive learning for novel class discovery. In: CVPR (2021) [7] Sun, Y., Shi, Z., Liang, Y., Li, Y.: When and how does known class help discover unknown ones? provable understanding through spectral analysis. In: ICML, vol. 202, pp. 33014–33043 (2023) [8] Li, Z., Otholt, J., Dai, B., Hu, D., Meinel, C., Yang, H.: A closer look at novel class discovery from the labeled set. In: NeurIPS 2022 Workshop on Distribution Shifts: Connecting Methods and Applications (2022) [9] Troisemaine, C., Flocon-Cholet, J., Gosselin, S., Vaton, S., Reiffers-Masson, A., Lemaire, V.: A method for discovering novel classes in tabular data. In: ICKG, pp. 265–274 (2022) [10] Caron, M., Touvron, H., Misra, I., Jegou, H., Mairal, J., Bojanowski, P., Joulin, A.: Emerging properties in self-supervised vision transformers. In: ICCV, pp. 1–21 (2021) [11] Vaze, S., Han, K., Vedaldi, A., Zisserman, A.: Generalized category discovery. In: CVPR, pp. 7492–7501 (2022) [12] Fei, Y., Zhao, Z., Yang, S., Zhao, B.: Xcon: Learning with experts for fine-grained category discovery. In: British Machine Vision Conference (BMVC) (2022) [13] Zhang, L., Qi, L., Yang, X., Qiao, H., Yang, M.-H., Liu, Z.: Automatically Discovering Novel Visual Categories with Self-supervised Prototype Learning (2022) [14] Chen, Y., Zhu, X., Li, W., Gong, S.: Semi-supervised learning under class distribution mismatch. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 34, pp. 3569–3576 (2020) [15] Guo, L.-Z., Zhang, Z.-Y., Jiang, Y., Li, Y.-F., Zhou, Z.-H.: Safe deep semisupervised learning for unseen-class unlabeled data. In: ICML, pp. 3897–3906 (2020) [16] Han, K., Vedaldi, A., Zisserman, A.: Learning to discover novel visual categories via deep transfer clustering. In: ICCV (2019) [17] Hsu, Y.-C., Lv, Z., Schlosser, J., Odom, P., Kira, Z.: Multi-class classification without multi-class labels. In: ICLR (2019) [18] Zheng, J., Li, W., Hong, J., Petersson, L., Barnes, N.: Towards open-set object detection and discovery. In: CVPR, pp. 3961–3970 (2022) [19] Yang, M., Zhu, Y., Yu, J., Wu, A., Deng, C.: Divide and conquer: Compositional experts for generalized novel class discovery. In: CVPR, pp. 14268–14277 (2022) [20] Cao, K., Brbic, M., Leskovec, J.: Open-world semi-supervised learning. In: ICLR (2022) [21] Sun, Y., Li, Y.: Opencon: Open-world contrastive learning. In: TMLR (2023) [22] Chi, H., Liu, F., Yang, W., Lan, L., Liu, T., Han, B., Niu, G., Zhou, M., Sugiyama, M.: Meta discovery: Learning to discover novel classes given very limited data. In: ICLR (2022) [23] Gidaris, S., Singh, P., Komodakis, N.: Unsupervised representation learning by predicting image rotations. In: ICLR (2018) [24] Arthur, D., Vassilvitskii, S.: K-means++ the advantages of careful seeding. In: ACM-SIAM SODA, pp. 1027–1035 (2007) [25] Luxburg, U.: A tutorial on spectral clustering. Stat. Comput. 17(4), 395–416 (2007) [26] Ng, A., Jordan, M., Weiss, Y.: On spectral clustering: Analysis and an algorithm. Advances in neural information processing systems 14 (2001) [27] Khan, A.A., Mohanty, S.K.: A fast spectral clustering technique using mst based proximity graph for diversified datasets. Information Sciences 609, 1113–1131 (2022) [28] Stuetzle, W.: Estimating the cluster tree of a density by analyzing the minimal spanning tree of a sample. Journal of classification 20(1), 25–47 (2003) [29] Ester, M., Kriegel, H.-P., Sander, J., Xu, X., et al.: A density-based algorithm for discovering clusters in large spatial databases with noise. In: Kdd, vol. 96, pp. 226–231 (1996) [30] Le, L., Patterson, A., White, M.: Supervised autoencoders: Improving generalization performance with unsupervised regularizers. Advances in neural information processing systems 31 (2018) [31] Zhao, B., Han, K.: Novel visual category discovery with dual ranking statistics and mutual knowledge distillation. In: Advances in Neural Information Processing Systems (2021) [32] Xie, J., Girshick, R., Farhadi, A.: Unsupervised deep embedding for clustering analysis. In: ICML, vol. 48, pp. 478–487 (2016) [33] Arbelaitz, O., Gurrutxaga, I., Muguerza, J., P´erez, J.M., Perona, I.: An extensive comparative study of cluster validity indices. Pattern recognition 46(1), 243–256 (2013) [34] Yang, M., Wang, L., Deng, C., Zhang, H.: Bootstrap your own prior: Towards distribution-agnostic novel class discovery. In: CVPR, pp. 3459–3468 (2023) [35] Anguita, D., Ghio, A., Oneto, L., Parra, X., Reyes-Ortiz, J.L.: A public domain dataset for human activity recognition using smartphones. In: ESANN (2013) [36] Frey, P.W., Slate, D.J.: Letter recognition using holland-style adaptive classifiers. Machine Learning 6, 161–182 (2005) [37] Dua, D., Graff, C.: UCI Machine Learning Repository (2017) [38] Kuhn, H.W., Yaw, B.: The hungarian method for the assignment problem. Naval Res. Logist. Quart, 83–97 (1955) [39] Von Luxburg, U., Williamson, R.C., Guyon, I.: Clustering: Science or art? In: ICML Workshop on Unsupervised and Transfer Learning, pp. 65–79 (2012) [40] Satopaa, V., Albrecht, J., Irwin, D., Raghavan, B.: Finding a” kneedle” in a haystack: Detecting knee points in system behavior. In: ICDCS Workshops, pp. 166–171 (2011). IEEE [41] Arvai, K.: kneed. Zenodo (2023). https://doi.org/10.5281/zenodo.7873825 This paper is available on arxiv under CC 4.0 license. Authors: (1) Troisemaine Colin, Department of Computer Science, IMT Atlantique, Brest, France., and Orange Labs, Lannion, France; (2) Reiffers-Masson Alexandre, Department of Computer Science, IMT Atlantique, Brest, France.; (3) Gosselin Stephane, Orange Labs, Lannion, France; (4) Lemaire Vincent, Orange Labs, Lannion, France; (5) Vaton Sandrine, Department of Computer Science, IMT Atlantique, Brest, France. Authors: Authors: (1) Troisemaine Colin, Department of Computer Science, IMT Atlantique, Brest, France., and Orange Labs, Lannion, France; (2) Reiffers-Masson Alexandre, Department of Computer Science, IMT Atlantique, Brest, France.; (3) Gosselin Stephane, Orange Labs, Lannion, France; (4) Lemaire Vincent, Orange Labs, Lannion, France; (5) Vaton Sandrine, Department of Computer Science, IMT Atlantique, Brest, France. Table of Links Abstract and Intro Abstract and Intro Related work Related work Approaches Approaches Hyperparameter optimization Hyperparameter optimization Estimating the number of novel classes Estimating the number of novel classes Full training procedure Full training procedure Experiments Experiments Conclusion Conclusion Declarations Declarations References References Appendix A: Additional result metrics Appendix A: Additional result metrics Appendix B: Hyperparameters Appendix B: Hyperparameters Appendix C: Cluster Validity Indices numerical results Appendix C: Cluster Validity Indices numerical results Appendix D: NCD k-means centroids convergence study Appendix D: NCD k-means centroids convergence study 8 Conclusion In this article, we have shown that in the NCD setting, unsupervised clustering algorithms can benefit from knowledge of the known classes and reliably improve their performance by implementing simple modifications. We have also introduced a novel NCD algorithm called PBN, which is characterized by its simplicity and low number of hyperparameters, which proved to be a decisive advantage under realistic conditions. In addition, we have proposed an adaptation of the k-fold cross-validation process to tune the hyperparameters of NCD methods without depending on the labels of the novel classes. Finally, we have demonstrated that the number of novel classes can be accurately estimated within the latent space of PBN. These two previous contributions have shown that the NCD problem can be solved in realistic situations where no prior knowledge of the novel classes is available during training. Declarations Funding Colin Troisemaine, Alexandre Reiffers-Masson, Stephane Gosselin, Vincent Lemaire and Sandrine Vaton received funding from Orange SA. Competing Interests Colin Troisemaine, St´ephane Gosselin and Vincent Lemaire received research support from Orange SA. Alexandre Reiffers-Masson and Sandrine Vaton received research support from IMT Atlantique. Ethics approval Not applicable. Consent to participate All authors have read and approved the final manuscript. Consent for publication Not applicable. Availability of data and materials All data used is this study are available publicly online. The datasets were extracted directly in the repositories available with the links in the corresponding section. Code availability The code for experiments is available at the following url: https://github.com/ PracticalNCD/ECMLPKDD2024. Authors’ contributions Colin Troisemaine, Alexandre Reiffers-Masson, St´ephane Gosselin, Vincent Lemaire and Sandrine Vaton contributed to the manuscript equally. References [1] Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., Fei-Fei, L.: Imagenet: A largescale hierarchical image database. In: CVPR, pp. 248–255 (2009) [2] Troisemaine, C., Lemaire, V., Gosselin, S., Reiffers-Masson, A., Flocon-Cholet, J., Vaton, S.: Novel class discovery: an introduction and key concepts. ArXiv (2023) [3] Hsu, Y.-C., Lv, Z., Kira, Z.: Learning to cluster in order to transfer across domains and tasks. In: ICLR (2018) [4] Han, K., Rebuffi, S.-A., Ehrhardt, S., Vedaldi, A., Zisserman, A.: Autonovel: Automatically discovering and learning novel visual categories. PAMI (2021) [5] Zhong, Z., Zhu, L., Luo, Z., Li, S., Yang, Y., Sebe, N.: Openmix: Reviving known knowledge for discovering novel visual categories in an open world. In: CVPR, pp. 9462–9470 (2021) [6] Zhong, Z., Fini, E., Roy, S., Luo, Z., Ricci, E., Sebe, N.: Neighborhood contrastive learning for novel class discovery. In: CVPR (2021) [7] Sun, Y., Shi, Z., Liang, Y., Li, Y.: When and how does known class help discover unknown ones? provable understanding through spectral analysis. In: ICML, vol. 202, pp. 33014–33043 (2023) [8] Li, Z., Otholt, J., Dai, B., Hu, D., Meinel, C., Yang, H.: A closer look at novel class discovery from the labeled set. In: NeurIPS 2022 Workshop on Distribution Shifts: Connecting Methods and Applications (2022) [9] Troisemaine, C., Flocon-Cholet, J., Gosselin, S., Vaton, S., Reiffers-Masson, A., Lemaire, V.: A method for discovering novel classes in tabular data. In: ICKG, pp. 265–274 (2022) [10] Caron, M., Touvron, H., Misra, I., Jegou, H., Mairal, J., Bojanowski, P., Joulin, A.: Emerging properties in self-supervised vision transformers. In: ICCV, pp. 1–21 (2021) [11] Vaze, S., Han, K., Vedaldi, A., Zisserman, A.: Generalized category discovery. In: CVPR, pp. 7492–7501 (2022) [12] Fei, Y., Zhao, Z., Yang, S., Zhao, B.: Xcon: Learning with experts for fine-grained category discovery. In: British Machine Vision Conference (BMVC) (2022) [13] Zhang, L., Qi, L., Yang, X., Qiao, H., Yang, M.-H., Liu, Z.: Automatically Discovering Novel Visual Categories with Self-supervised Prototype Learning (2022) [14] Chen, Y., Zhu, X., Li, W., Gong, S.: Semi-supervised learning under class distribution mismatch. In: Proceedings of the AAAI Conference on Artificial Intelligence, vol. 34, pp. 3569–3576 (2020) [15] Guo, L.-Z., Zhang, Z.-Y., Jiang, Y., Li, Y.-F., Zhou, Z.-H.: Safe deep semisupervised learning for unseen-class unlabeled data. In: ICML, pp. 3897–3906 (2020) [16] Han, K., Vedaldi, A., Zisserman, A.: Learning to discover novel visual categories via deep transfer clustering. In: ICCV (2019) [17] Hsu, Y.-C., Lv, Z., Schlosser, J., Odom, P., Kira, Z.: Multi-class classification without multi-class labels. In: ICLR (2019) [18] Zheng, J., Li, W., Hong, J., Petersson, L., Barnes, N.: Towards open-set object detection and discovery. In: CVPR, pp. 3961–3970 (2022) [19] Yang, M., Zhu, Y., Yu, J., Wu, A., Deng, C.: Divide and conquer: Compositional experts for generalized novel class discovery. In: CVPR, pp. 14268–14277 (2022) [20] Cao, K., Brbic, M., Leskovec, J.: Open-world semi-supervised learning. In: ICLR (2022) [21] Sun, Y., Li, Y.: Opencon: Open-world contrastive learning. In: TMLR (2023) [22] Chi, H., Liu, F., Yang, W., Lan, L., Liu, T., Han, B., Niu, G., Zhou, M., Sugiyama, M.: Meta discovery: Learning to discover novel classes given very limited data. In: ICLR (2022) [23] Gidaris, S., Singh, P., Komodakis, N.: Unsupervised representation learning by predicting image rotations. In: ICLR (2018) [24] Arthur, D., Vassilvitskii, S.: K-means++ the advantages of careful seeding. In: ACM-SIAM SODA, pp. 1027–1035 (2007) [25] Luxburg, U.: A tutorial on spectral clustering. Stat. Comput. 17(4), 395–416 (2007) [26] Ng, A., Jordan, M., Weiss, Y.: On spectral clustering: Analysis and an algorithm. Advances in neural information processing systems 14 (2001) [27] Khan, A.A., Mohanty, S.K.: A fast spectral clustering technique using mst based proximity graph for diversified datasets. Information Sciences 609, 1113–1131 (2022) [28] Stuetzle, W.: Estimating the cluster tree of a density by analyzing the minimal spanning tree of a sample. Journal of classification 20(1), 25–47 (2003) [29] Ester, M., Kriegel, H.-P., Sander, J., Xu, X., et al.: A density-based algorithm for discovering clusters in large spatial databases with noise. In: Kdd, vol. 96, pp. 226–231 (1996) [30] Le, L., Patterson, A., White, M.: Supervised autoencoders: Improving generalization performance with unsupervised regularizers. Advances in neural information processing systems 31 (2018) [31] Zhao, B., Han, K.: Novel visual category discovery with dual ranking statistics and mutual knowledge distillation. In: Advances in Neural Information Processing Systems (2021) [32] Xie, J., Girshick, R., Farhadi, A.: Unsupervised deep embedding for clustering analysis. In: ICML, vol. 48, pp. 478–487 (2016) [33] Arbelaitz, O., Gurrutxaga, I., Muguerza, J., P´erez, J.M., Perona, I.: An extensive comparative study of cluster validity indices. Pattern recognition 46(1), 243–256 (2013) [34] Yang, M., Wang, L., Deng, C., Zhang, H.: Bootstrap your own prior: Towards distribution-agnostic novel class discovery. In: CVPR, pp. 3459–3468 (2023) [35] Anguita, D., Ghio, A., Oneto, L., Parra, X., Reyes-Ortiz, J.L.: A public domain dataset for human activity recognition using smartphones. In: ESANN (2013) [36] Frey, P.W., Slate, D.J.: Letter recognition using holland-style adaptive classifiers. Machine Learning 6, 161–182 (2005) [37] Dua, D., Graff, C.: UCI Machine Learning Repository (2017) [38] Kuhn, H.W., Yaw, B.: The hungarian method for the assignment problem. Naval Res. Logist. Quart, 83–97 (1955) [39] Von Luxburg, U., Williamson, R.C., Guyon, I.: Clustering: Science or art? In: ICML Workshop on Unsupervised and Transfer Learning, pp. 65–79 (2012) [40] Satopaa, V., Albrecht, J., Irwin, D., Raghavan, B.: Finding a” kneedle” in a haystack: Detecting knee points in system behavior. In: ICDCS Workshops, pp. 166–171 (2011). IEEE [41] Arvai, K.: kneed. Zenodo (2023). https://doi.org/10.5281/zenodo.7873825 This paper is available on arxiv under CC 4.0 license. This paper is available on arxiv under CC 4.0 license. available on arxiv

Stories, News & Mentions about HOLLAR

Hollar’s Stories on HackerNoon

Hollar's News & Mentions