Deccan Herald: Flim: A bot that can guess the exact movie you are thinking of.”Within a year, it already has the world’s largest database of film-related images. Flim’s unique algorithm scans through more than three lakh high-definition images from movies, documentaries, anime, advertising and music videos. Its bot doesn’t just detect broader aspects like colour palette, genre and aspect ratio, but also identify details like dog-walkers, ham sandwiches, red-lipped Asian women, animated metropolises and any other clue.” Three lakh is 300,000. I suspect it’s going to take a lot of experimenting to get a grip on this search engine.
Platform launch: Qoruz launches India’s first comprehensive influencer search engine (The Tech Panda)
The Tech Panda: Platform launch: Qoruz launches India’s first comprehensive influencer search engine. “The search engine, termed Qoruz Search, gives access to India’s largest database of hundreds and thousands of influencers, including celebrities, macro, micro and nano influencers for free across Instagram, YouTube, Twitter, Facebook, and blogs. Advanced features like filters, detailed analytics, insights and campaign planning tools will be available for a small license fee.”
PR Newswire: accessiBe, The Web Accessibility Market Leader, Announces New Search Engine accessFind to Help People with Disabilities Find Accessible Websites (PRESS RELEASE)
PR Newswire: accessiBe, The Web Accessibility Market Leader, Announces New Search Engine accessFind to Help People with Disabilities Find Accessible Websites (PRESS RELEASE). “Through partnerships with organizations including United Spinal Association, Columbia Lighthouse For The Blind (CLB), The Viscardi Center, The IMAGE Center of Maryland, Earle Baum Center of the Blind, Determined2Heal, Senspoint, and others, accessFind will be the first search engine designed to enhance how people with disabilities navigate the internet through the creation of an accessible-friendly index of websites.”
Children’s Hospital of Philadelphia: CHOP Researchers Create Search Engine Linking Diseases and Observable Traits
Children’s Hospital of Philadelphia: CHOP Researchers Create Search Engine Linking Diseases and Observable Traits . “A patient comes to a doctor with a constellation of symptoms and physical traits that, at a first glance, do not lead to an obvious diagnosis. Perhaps the patient has facial abnormalities and cardiac symptoms, as well as other issues with the skeleton and joints. The provider suspects these symptoms relate to a disease that is due to an underlying genetic mutation but is unsure where to look. Enter PhenCards, a new data resource and search engine created by researchers at Children’s Hospital of Philadelphia (CHOP) that links existing biomedical knowledge with observable human traits, also known as phenotypes.”
MIT Technology Review: Language models like GPT-3 could herald a new type of search engine. “…a team of Google researchers has published a proposal for a radical redesign that throws out the ranking approach and replaces it with a single large AI language model—a future version of BERT or GPT-3. The idea is that instead of searching for information in a vast list of web pages, users would ask questions and have a language model trained on those pages answer them directly. The approach could change not only how search engines work, but how we interact with them.” Gee, like natural language searching? Like MIT sued Ask Jeeves over?
PR Newswire: Clearity Health, A Crowdsourced Search Engine For Local Healthcare Prices, Launches And Calls For Medical Bills (PRESS RELEASE)
PR Newswire: Clearity Health, A Crowdsourced Search Engine For Local Healthcare Prices, Launches And Calls For Medical Bills (PRESS RELEASE). “Clearity Health, a new organization created by successful local founders and medical doctors, launches today in Austin. Its namesake healthcare cost search engine will show users exactly what other patients with similar conditions have paid at local hospitals and medical offices, including information about what services may be required.” There’s a waiting list for what looks like a closed beta. Not clear if this is Austin-only, Texas-only, or nationwide.
FedScoop: USPTO chief information officer most excited about new search algorithms . “New search algorithms for relevant prior art most excite the U.S. Patent and Trademark Office’s CIO right now. USPTO created the machine-learning algorithms to increase the speed at which patents are examined by importing relevant prior art — all information on its claim of originality — into pending applications sent to art units, said Jamie Holcombe.”
Spotted via Reddit: ZodiacNFT. From the About page: “Zodiac is a visual discovery engine for NFTs. Combined with a set of proprietary AI modules, Zodiac enables users to search for and discover the best content on blockchain.” It’s a visual search engine for NFTs. I put in a picture of my mother’s dog and got lots of NFTs of dogs and cats. I put in a Norma Shearer meme and got pictures of Marilyn Monroe.
News18: ‘Same Energy’ is a Visual Search Engine Which Finds Similar Images to Match the ‘Vibe’ Perfectly
News18: ‘Same Energy’ is a Visual Search Engine Which Finds Similar Images to Match the ‘Vibe’ Perfectly. “Same Energy is a new web tool (still in development) that’s perfect for anyone looking for visual inspiration – It’s an AI-powered visual search engine that provides a fast and simple experience for exploring visually similar photos.” I played with it a little. Loved its wicked-fast response and the way the results filled the screen.
Washington Post: How Google is hurting local news. “As an example, we examined the proportion of local and national search results for ‘early voting’ that Google News returned Oct. 26, before the 2020 election. Certainly at least some people searching for that term hoped to get information about local or state early voting information. However, only 20 percent of the top 10 returned searches were from local outlets. If readers kept scrolling past the 20th result, they would eventually find local outlets. But national outlets were the most common sources returned in the top results — and therefore more likely to be clicked. In the top two results, national outlets’ stories appeared 56 percent of the time. Google News’s top pick was a national outlet almost 74 percent of the time.”
US Copyright Office: Copyright Office Launches Copyright Public Records System Pilot. “Today, the U.S. Copyright Office launched a new Copyright Public Records System (CPRS) pilot to the public. The new portal will provide access to the same copyright records for both registration and recordation data that exist in the Copyright Public Catalog but with enhanced search capabilities and improved interfaces for internal and external users.”