Google पेटेंट आर्काइव

Google और Microsoft search patents का एक curated archive organized inventor या research lead द्वारा जिसने उन्हें shipped किया। प्रत्येक track एक engineer को उनकी patent lineage के माध्यम से follow करता है ताकि reader देख सके कि एक research thread continuations, divisional applications, और grants में कैसे evolved हुआ। Patents को annotated किया जाता है ranking surface के साथ जिसे वे touch करते हैं: query understanding, ranking, retrieval, behavioral signals, knowledge graph, या AI search।

Patent tracks by inventor

• 65 Google Patents to Help You Understand How Search Engines Work! (69 patents) — Google द्वारा filed foundational और advanced patents का एक curated collection।
• Steven Baker — Google Search Patents (49 patents) — 49 granted search-engine patents plus 2 published applications Google researcher Steven D. Baker द्वारा, covering query understanding (synonyms, n-gram, geographic, cross-language), query refinement, meaningful stopword detection, answer passage scoring (featured snippets), query–document similarity, list co-occurrence, और embedding-based personalized search।
• Amit Singhal — Google Search Patents (22 patents) — 22 search-engine patents Amit Singhal द्वारा, former Google SVP of Search (2000-2016), covering interleaving और result composition, query rewriting, semantic query understanding, link-quality scoring, location-aware ranking, language और country biasing, commercial-intent detection, और advertisement serving। 8-patent meaningful-stopwords family Steven Baker के साथ co-invented है और duplicate content से बचने के लिए Steven Baker section में documented है।
• Prabhakar Raghavan — Google Search Patents (16 patents) — 16 search/IR patents Prabhakar Raghavan द्वारा, head of Google Search 2020-2024 (previously Yahoo VP और IBM Almaden research scientist), covering personalized PageRank, web community detection, adaptive ranking, multi-engine result merging, hierarchical taxonomy, subspace clustering, information filtering, customizable navigation, user-generated content surfacing, location-quality signals from travel patterns, और temporal metadata analysis। उनके IBM, Yahoo, और Google research को span करता है।
• Nitin Gupta — Google Search Patents (32 patents) — 32 search/SEO patents Nitin Gupta द्वारा (Google Search Quality engineer) covering query suggestions और templates, reranking और personalization, answer-generating neural networks, content-knowledge generation, latest 2026 generative-search retrieval patent AI Overviews era के लिए, और एक direct SEO patent domain names और websites पर। Final 5 patents Steven Baker के साथ co-invented हैं और duplicate content से बचने के लिए उनके section में documented हैं।
• Navneet Panda — Google Search Patents (15 patents) — 15 search-quality patents Navneet Panda द्वारा, Google engineer जिसका content-quality work Panda algorithm update (2011) के लिए basis बन गया। Covers site quality scoring, predicting site quality behavioral data accumulate होने से पहले, ranking search results combined relevance और site-quality signals द्वारा, website duration performance, locally significant queries, query revision, और image-concept learning।
• Anand Shukla — Google Search Patents (16 patents) — 16 search/IR patents Anand Shukla द्वारा, Google Search Quality engineer, including cutting-edge 2024 "Search with stateful chat" patent जो classical search को LLM-based conversational search से bridge करता है। Also covers in-depth article surfacing, query pattern matching, query topic mapping, content channel curation, और enhanced-search-feed family। Final 3 patents Steven Baker के साथ co-invented हैं personalized-feed embeddings work पर और उनके section में documented हैं।
• Krishna Bharat, Google Search Patents (48 patents) — 48 search-engine patents Krishna Bharat द्वारा, founding inventor of Google News और creator of Hilltop algorithm। Covers Hilltop link-analysis algorithm, query rewriting with entity detection, semantic units in queries, query और search augmentation, document ranking by usage statistics, image selection for news search, editorial-opinion ranking, search engine coverage estimation, query-specific bookmarking, और entity-aware ad rendering। 2001 से 2024 तक span करता है।
• Srinivasan Venkatachary, Google Search Patents (25 patents) — 26 search-engineering patents Srinivasan Venkatachary द्वारा, co-inventor on Generative Summaries for Search Results patent family जो AI Overviews और Search Generative Experience को underpin करता है। Also covers enhanced search feeds based on user interests, candidate-answer-passage scoring, context-scoring adjustments for answer passages, implicit question query identification, keyword-based conversational voice search, query generation via document structural similarity, resource identification from organic और structured content, content channel curation, और emoji classifier। 2013 से 2025 तक span करता है।
• Jeromy William Henry, Google Search Patents (25 patents) — 25 search-engine patents Jeromy William Henry द्वारा covering Knowledge Panels (entity card search results के बगल में), Structured Entity Information Pages (standalone entity-detail surface), Personalized Entity Information Pages, SERP image sizing, और music-search secondary link presentation। उनके corpus में हर patent SEO-relevant है। 2013 से 2025 तक span करता है।
• Ramanathan V. Guha, Google Search Patents (33 patents) — 33 search-engine patents Ramanathan V. Guha द्वारा, co-founder of Schema.org और lead inventor on Google Custom Search Engine। Covers Programmable Search Engine family, trust-based search ranking, spam detection for programmable contexts, social-signal propagation, classifying search results, customized web summaries, filtering by user annotations, contextual searching by query intersections, query identification, और generating-tasks UI family। 2003 से 2019 तक span करता है।
• Marc Najork, Google Search Patents (25 patents) — 25 search/IR patents Marc Najork द्वारा spanning उनके Microsoft Research era (web crawler architecture, spam-resistant ranking, content evaluation, hyperlink databases) और उनके Google era (form information extraction, ML-driven semantic similarity, active learning, document activity logs for relevance training)। Includes एक foundational cross-listed patent 65 Google Patents collection के साथ document activity logs पर। 2005 से 2025 तक span करता है।
• Jeffrey Dean, Google Search Patents (78 patents) — 78 search और IR-infrastructure patents Jeffrey Dean द्वारा, Google Senior Fellow और Chief Scientist, co-founder of Google Brain, और architect of MapReduce और BigTable। Includes foundational Document Scoring families (Query Analysis, Content Update, Inception Date, Link-Based Criteria), Historical Data Patent (cross-listed with 65gp), Anchor Tag Indexing, Content Snippets via Tokenspace, near-duplicate और mirrored-host detection, topic-drift prevention, distributed ML models, और foundational MapReduce और BigTable infrastructure जो web-scale search को possible बनाता है। 2000 से 2025 तक span करता है।
• Paul Haahr, Google Search Patents (63 patents) — 63 Google search और ranking patents Paul Haahr द्वारा, long-tenured Google search ranking engineer। Covers canonical link-quality patents (Bharat/Singhal/Haahr "Determining quality of linked documents"), document-classification ranking, node-independence anti-PBN signal, rumored quality-measure patent (cross-listed with 65gp pat-48), multi-stage query processing pipeline, framework for evaluating web search scoring functions (उनके SMX talk के documentary back-stop), preferred sites, reference contexts, synthetic descriptive text for images, और multi-family query refinement और stopword-detection clusters। Co-authored with Singhal, Dean, Bharat, Cutts, Henzinger, Simon Tong, और Acharya। 2008 से 2019 तक span करता है।
• Hyung-Jin Kim, Google Search Patents (23 patents) — 23 Google search और ranking patents Hyung-Jin Kim द्वारा। Lead inventor on Navboost / implicit-feedback ranking family (US 8,661,029 + 5 continuations through US 11,816,114 active 2023 IP), system 2024 DOJ Google antitrust trial पर surfaced। Also covers presentation-bias model, temporal score adjustments, document-change ranking, user cohort grouping, similar-query borrowing, variant generalized queries, propagating query classifications, और cross-listed Quality Measure patent (with Paul Haahr, 65gp pat-48)। Co-authored with Simon Tong, Noam Shazeer, Michelangelo Diligenti, Adrian Corduneanu, Henele Adams, Andrei Lopatenko। Filings 2006-2021।
• Pandu Nayak, Google Search Patents (13 patents) — 13 Google search patents P. Pandurang Nayak द्वारा, Google Chief Scientist for Search। Patent record query-revision और query-understanding stack को 2003-2015 से cover करता है: integration of multiple query-revision models, confidence-scored revision, known-highly-ranked-queries borrowing, concept-context substitution, phrase-restricted synonyms, click-driven synonym identification, acronym expansion। Co-authored with David R. Bailey, Ben Gomes, Trystan Upstill, Kedar Dhamdhere, Thomas Strohmann। उनके neural-era contributions (RankBrain, BERT, MUM, DOJ testimony) intentionally unpatented हैं और blog posts और trial transcripts में live हैं।
• Daniel Egnor, Google Local Search Patents (20 patents) — 20 Google local-search और geographic-relevance patents Daniel Egnor द्वारा, architect of Google Maps / Local Pack ranking layer। Lead inventor on US 8,046,371 "Scoring local search results based on location prominence" - patent जो "Prominence" factor को Google के official Local Search Ranking documentation में define करता है। Also covers location-sensitivity ranking (Relevance factor, with Amit Singhal), geographical-relevance indexing, ambiguous-geographic-reference resolution, authoritative-document identification, business-listing categorization, और GIS ambiguous-search processing (with Keyhole founder John Hanke)। Co-authored with Singhal, Haahr, Greenfield, John Hanke। Filings 2003-2019।
• Yossi Matias, Google SERP Features & Assistant Patents (47 patents) — 47 Google patents Yossi Matias द्वारा, Google VP Engineering। Lead inventor on foundational People-Also-Ask patent US 9,213,748 "Generating related questions for search queries"। Built Autocomplete (2004), Trends, weather/sports/dictionary OneBoxes। Covers People-Also-Ask family, Knowledge Panel entity scaffolding (grouping related entities, central-entity identification, fact estimation, textual entity attributes, entity-quote search), Trends/analytics layer (campaign-competitive analysis, ranking graphical visualizations), Duplex automated-call infrastructure, personalized chatbot और LLM-output entailment patents जो Google Assistant के 2024-2026 generative era को underpin करते हैं, और contextual / third-party SERP-features। Co-authored with Yaniv Leviathan (Duplex), Gal Chechik, Ziv Bar-Yossef, Eyal Segalis, Avinatan Hassidim। Filings 2013-2026।
• Andrei Broder, Cross-Vendor Search Patents (25 patents) — 25 search-engine और IR patents Andrei Broder द्वारा, cross-vendor inventor जिसने DEC/AltaVista में near-duplicate detection को shape किया, foundational CAPTCHA को Compaq में co-invented किया, AltaVista web-page ranking system को built किया, IBM में web-page-decay signal को authored किया, और Yahoo में ad-matching/SERP-features/CAPTCHA-evolution infrastructure को shape किया। Lead inventor on foundational MinHash/shingling patent (US 6,349,296), fingerprint-collision-probability technique (US 5,974,481), document-resemblance shingling patent (US 6,230,155), और foundational CAPTCHA patent (US 6,195,698)। 1999 से 2024 तक span करता है।
• Monika Henzinger, Google Search Patents (50 patents) — 98 search और IR patents Monika Henzinger द्वारा, first Director of Research at Google। Co-inventor on foundational near-duplicate-page-detection patent (US 6,138,113, with Dean) और document-scoring families का extensive set (Query Analysis, Content Update, Inception Date, Link-Based Criteria, Historical Data - all cross-listed with Dean's canonical articles)। Independent contributions: detecting-duplicate-files family with William Pugh, AltaVista connectivity-server, connectivity-और-content-ranking patents, document-freshness determination, semantic-distance ranking, query-semantic-information ranking, anchor-text cross-language IR, in-context searching, hypertext-browser assistant, और usage-statistics-driven document retrieval। 1997 से 2017 तक span करता है।
• Trystan Upstill, Google Site-Quality & HCU-Era Search Patents (28 patents) — 28 Google search और ranking patents Trystan Upstill द्वारा। Co-inventor on Panda patent (US 9,135,307 = 65gp pat-39) और lead inventor on HCU-era follow-up US App 2019/0155948 "Re-ranking resources based on categorical quality" - patent जो post-Panda site-quality lineage को anchor करता है। Also covers authoritative-results discovery, search-result ranking और re-scoring, affiliated-domains detection, resource-attribute extraction from site address, local search और POI identification, synonyms और query rewriting, navigational-intent resources, language identification from link context, और recent LLM-era UI components patent। Filings 2012-2026।
• Susan Dumais, Microsoft IR & Search Patents (18 patents) — 61 search और IR patents Susan Dumais द्वारा, Microsoft Technical Fellow और Gerard Salton Award winner। Co-inventor on foundational Latent Semantic Indexing patent (US 4,839,853, Bellcore 1989) - conceptual ancestor of हर dense-embedding retrieval system। Microsoft Research पर उनका काम automated SERP satisfaction measurement, preference-judgment click models, activity-based-context ranking, time-aware ranking with temporal dynamics, personalized navigation और search results, per-user domain-expertise determination, web-page-change × revisitation freshness, implicit device-related query reformulation, authority ranking को cover करता है। 1989 से 2026 तक span करता है। Co-authors में Jaime Teevan, Eric Horvitz, Ryen White, Adam Fourney, Joshua Goodman, Eric Brill शामिल हैं।
• Christopher Burges, Microsoft Research Learning-to-Rank Patents (16 patents) — ~30 unique US inventions (84-entry WIPO corpus) Christopher J. C. Burges द्वारा, Microsoft Research scientist जिसने RankNet/LambdaRank/LambdaMART learning-to-rank lineage को invented किया। RankNet (US 7,689,615) pairwise neural ranker है, LambdaRank (US 7,617,164) arbitrary-cost extension है जो NDCG को optimize करता है, LambdaMART (US 12/032,697) gradient-boosted ensemble है जो Yahoo LTR Challenge 2010 को जीता। Also covers behavioral-variability adaptation, web-page multi-graph analysis, link-spam smoothing, margin-tree boosting, k-NN probability estimation। 2001 से 2019 तक span करता है। Co-authors में Robert Ragno (LambdaRank), Irina Matveeva, Timo Burkard शामिल हैं।
• Eric Brill, Microsoft Research Search & NLP Patents (20 patents) — 20 Microsoft Research patents Eric Brill द्वारा, NLP और search scientist query speller patents के लिए known और noisy-channel correction model। Lead inventor on US 7,254,774 "Systems and methods for improved spell checking" - foundational noisy-channel query speller patent। Also covers string-to-string spell-correction transformations (US 7,290,209 / 7,366,983), behavioral-variability accounting in web search (US 7,743,047), popularity-data ranking, mining web search user behavior, page-biased search, cost-benefit Q&A composition, और user-intent discovery। Filings 2005-2010।
• Tomas Mikolov, Google word2vec Patents (2 patents) — 2 foundational Google patents Tomas Mikolov द्वारा word2vec पर covering - continuous word-embedding learning model जो embeddings era को defined करता है। Co-invented with Kai Chen, Greg Corrado, और Jeff Dean। US 9,037,464 original grant है, US 9,740,680 continuation है broader claim scope के साथ। दोनों patents एक ही disclosure (application 13/841,640) share करते हैं और एक साथ संपूर्ण word2vec patent family को constitute करते हैं। Conceptually Dean के section (Dean co-inventor है और broader Brain/distributed-models infrastructure पर काम करता है) के साथ cross-listed है।
• Noam Shazeer, Google Transformer, MoE & Search Patents (18 patents) — ~46 captured Google patents (~120-150 का full portfolio estimated) Noam Shazeer द्वारा। Co-inventor on foundational Transformer attention architecture (US 10,452,978 with Vaswani, Polosukhin, Uszkoreit, Jones, Gomez, Kaiser, Parmar), Sparsely-Gated Mixture-of-Experts scaling approach (US 11,769,055 with Dean, Hinton, Le, Mirhoseini), Switch Transformer (US 12,093,829), और Navboost implicit-feedback ranking family (US 8,661,029 with Kim/Tong/Diligenti - cross-listed)। Also covers LLM-in-assistant response generation, distributed tensor computations (GSPMD/Pathways), और 2008-2015 pre-Transformer large-scale ML ranking infrastructure। 2008 से 2025 तक span करता है।
• Ashish Vaswani, Google Transformer, Attention-Based Vision & Fast Decoding Patents (6 patents) — 4 captured new canonical articles plus 2 cross-listings Noam Shazeer section से। Lead author of foundational Transformer patent (US 10,452,978, cross-listed)। First inventor on local-self-attention computer vision patent (US 20210390410, HaloNet) जो attention को long sequences और images तक scale करता है। Co-inventor on MultiModel (US 20200089755 - unified multi-task multi-modal architecture जो MUM और Gemini का structural ancestor है), fully-attentional computer vision system (US 20250292560, ViT lineage), और fast-decoding-with-discrete-latents method (EP 3732627) जो LLM inference को search-scale latency पर economically viable बनाता है। 2017 से 2025 तक span करता है।
• Jakob Uszkoreit, Google Transformer Architect, Universal Transformer & Vision Transformer Patents (6 patents) — 3 new canonical articles plus 3 cross-listings Vaswani और Shazeer sections से। Co-author of "Attention Is All You Need" (US 10,452,978, cross-listed)। Co-inventor on Universal Transformer (US 10,740,433, adaptive-depth Transformer with per-token Adaptive Computation Time), Decomposable Attention NLI model (EMNLP 2016, load-bearing pre-Transformer proof जो attend-compare-aggregate को recurrence से replace करता है), और Vision Transformer (ViT, ICLR 2021, patch-tokenization architecture जो modern visual search और multimodal grounding को anchor करता है)। Cross-listings original Transformer, MultiModel (unified multi-task multi-modal ancestor of MUM/Gemini), और Image Transformer को cover करते हैं। 2016 से 2021 तक span करता है।
• Simon Tong, Google Click-Driven Ranking, Population Signals & Anti-Spam Patents (9 patents) — 5 new canonical articles plus 4 cross-listings Kim, Shazeer, और 65 Google Patents sections से। Tong co-inventor है Marc Pearson और Sergey Brin के साथ related-query ranking patent पर (US 7,505,964), Pearson के साथ population-information ranking patent पर (US 7,454,417), country biasing पर solo (US 20040254932), Bem/Harik/Levenberg/Shazeer के साथ pre-Transformer large-scale ML ranking infrastructure पर (US 7,222,127), और Ghemawat/Piscitello/Cutts के साथ user-document-removal patent पर (US 8,417,697, structural ancestor of Personal Blocklist और crowd-sourced spam signals)। Cross-listings Navboost implicit-feedback family, CTR-as-ranking-factor patent, historical-data patent, और large-dataset-ranking continuation को cover करते हैं। 2003 से 2017+ तक span करता है।
• Michelangelo Diligenti, Google Click-Data Quality & Implicit Feedback Patents (5 patents) — 2 new canonical articles plus 3 cross-listings Kim और 65 Google Patents sections से। Diligenti inventor है Detecting Click Spam patent (US 8,694,374, attribute-deviance anomaly detection feeding ranking signal) पर और Click Model That Accounts for User Intent पर (US 20120143789, intent-conditional click weighting)। उनके Navboost / presentation-bias / CTR-as-ranking-factor co-inventorships (cross-listed) implicit-feedback ranking family में उसे anchor करते हैं। Portfolio click-data quality पर focus करता है, raw user behavior और ranking signal के बीच layer जिसे यह feeds करता है। 2007 से 2019+ तक span करता है।
• Jaime Teevan, Microsoft Personalized Search & Search-Task Patents (5 patents) — 4 new canonical articles plus 1 cross-listing Dumais section से। Teevan co-inventor है Dumais और Horvitz के साथ foundational personalized-search patent पर (US 7,693,818) जो global candidates को per-user profile vector के विरुद्ध re-rank करता है। Inventor on search-task identification (US 8,326,824, cross-session query clustering into multi-step tasks), on re-finding vs new-finding classification (US 8,756,219, stabilizing prior clicked results for return visits), और on microtask search decomposition (US 10,062,103, decomposing high-level goals into resumable micro-steps)। Cross-listing personalized-navigation को (Dumais canonical) cover करता है। 2003 से 2018+ तक span करता है।
• Eric Horvitz, Microsoft Decision-Theoretic Search & Attention-Aware Information Delivery Patents (3 patents) — 3 new canonical articles Eric Horvitz द्वारा। Horvitz inventor है decision-theoretic ranking under uncertainty पर (US 7,124,129, relevance as probability distribution with expected-utility selection), cost-benefit reasoning in search-result presentation पर (US 7,194,464, mechanical basis for featured snippets और AI Overviews), और attention-aware proactive information delivery पर (US 6,934,917, zero-query surfaces lineage behind Discover, Now cards, और AI assistants)। 2002 से 2006+ तक span करता है।
• Ryen White, Microsoft Search Behavior & Search-Trails Patents (3 patents) — 3 new canonical articles Ryen White द्वारा। White inventor है search-trail / successful-path mining पर (US 7,904,440, ranking destinations जो end successful cross-user trails), exploratory-versus-lookup query phase classification पर (US 8,041,711, serving phase-tuned ranking और result formats), और cross-session task continuation पर (US 9,116,996, stitching queries across sessions और devices into one continuing task)। 2007 से 2015+ तक span करता है।
• Azalia Mirhoseini, Google ML Infrastructure & Chip-Design Patents (3 patents) — 2 new canonical articles plus 1 cross-listing Shazeer section से। Mirhoseini co-inventor है RL-based chip floorplan placement patent पर (US 11,475,278, Nature 2021 paper जो TPU floorplans को hours में weeks की बजाय design करता है) और device-placement-for-distributed-ML patent पर (arXiv 1706.04972, RL controller जो learns करता है कि कैसे large models को GPUs/TPUs में split करें training time को minimize करने के लिए)। दोनों compute-cost economics में back feed करते हैं जो Google को Transformer ranking को search-scale पर serve करने देते हैं। Cross-listing foundational MoE patent को (Shazeer canonical) cover करता है। 2017 से 2021+ तक span करता है।
• Arnold Overwijk, Microsoft Site Topical Authority Patent (1 patents) — 1 canonical article Arnold Overwijk द्वारा। Overwijk co-inventor है Li Xiong, Chuan Hu, और Junaid Ahmed के साथ Site Topical Authority patent पर (US 20210004416, formula A(s,t) = SR(s) * C(s) * T(s,t)^2 जो explicitly quantify करता है per-site per-topic authority global site rank, click signal, और squared per-topic page share को combining द्वारा)। Modern patent literature में topical authority का सबसे precise mathematical formulation। Conceptually Procopio (per-author per-topic authority) और Agent Rank (author-level reputation, 65gp pat-47) के साथ paired है।
• Michael Procopio, Google Per-Author Per-Topic Authority Patent (1 patents) — 1 canonical article Michael Procopio द्वारा। Procopio named inventor है Topic Authority patent पर (US 8,458,196)। System compute करता है per (author, topic) authority signatures: प्रत्येक document d के लिए जिसे author ने contribute किया, sum करता है (authorship% * topic_weight), author के सभी documents को aggregate करते हुए एक per-author per-topic expertise score produce करने के लिए। "यह person इस topic पर एक authority है" का mathematical quantification, structurally "इस author है expertise-level signal का foundation E-A-T पर।" Connect करता है Agent Rank (US 7,565,358, 65gp pat-47, earlier Google author-reputation patent) को और Overwijk के site-topical-authority (US 20210004416) को एक layered authority model बनाने के लिए।
• Ori Allon, Google Orion Query-Refinement Patent (1 patents) — 1 canonical article Ori Allon द्वारा। Allon co-inventor है Orion query-refinement patent पर (US 8,392,443) Ugo Di Girolamo, Tomer Shmiel, Alexandre Petcherski, और Tzvika Hartman के साथ। Orion algorithm Allon के PhD-thesis work से originate किया UNSW पर और Google द्वारा 2006 में acquired किया गया। Patent corpus-mined query refinements generate करता है: एक query दिया, phrase patterns को documents से extract करता है जो query को satisfy करते हैं ranked refinement candidates produce करने के लिए। "Searches related to," extended SERP snippets, और People-Also-Ask की lineage के पीछे mechanism। Granted 2013।
• Corin Anderson, Google Reasonable Surfer (Foundational) Patent (2 patents) — 1 new canonical article plus implicit cross-listings 65 Google Patents collection के लिए। Anderson FIRST inventor है foundational Reasonable Surfer patent पर (US 7,716,225, with Jeff Dean और Alexis Battle), 2004-filed grant जो non-uniform link-click probability को per-link features (anchor text, font size, position, surrounding context) से weighted introduce करता है। Continuations (US 8,117,209 in 65gp pat-45, US 9,305,099 in 2016) एक ही model को extend करते हैं। Mechanism एक ही page पर links में differentiated PageRank flow के लिए।
• Ashok Popat, Google OCR & Image-Text Indexing Patents (1 patents) — 1 canonical article Ashok Popat द्वारा covering उनके translation-inspired OCR research (Popat et al., 2009 onward, US 8,675,012 patent family for selective display of OCR'd text)। Methodology OCR को statistical machine translation के रूप में reframe करता है: image features source channel के रूप में, character sequences target के रूप में, joint decoding with channel और language models transcription quality को maximize करता है। Google Books indexing, Lens text capture, Drive PDF search, और image-bound-text indexing pipeline को power करता है हर image और infographic के पीछे web पर।

यह organization क्यों

Patent law intellectual property को inventor और assignee द्वारा cluster करता है, topic द्वारा नहीं। Inventor द्वारा grouping narrative arc को recover करता है: Amit Singhal का ranking का दशक, Pandu Nayak का BERT-era query handling, Krishna Bharat का news ranking lineage। प्रत्येक track एक research biography की तरह read करता है जो claim language में written है।