My experience spans technical AI leadership with expertise in founding, growing and managing ML engineering, science and product teams. My work involves research and industry-wide practical applications related to the fields of machine learning, natural language processing (NLP), computer vision, data mining and computational decipherment (cracking codes with computers). This has led to scientific advances and industry-wide innovations in conversational AI, on-device machine learning for IoT devices, large-scale unsupervised and semi-supervised methods and their applications to structured prediction problems in NLP, image recognition, information extraction, multi-modal learning for language/vision, user modeling in social media, graph optimization algorithms for summarizing noisy data, computational decipherment and computational advertising.

My work in the Press


• Neural Structured Learning in TensorFlow
  VentureBeat, Medium, SiliconANGLE
  Check out: TensorFlow package, GitHub, TensorFlow World Conference
  
  
• On-device Neural Networks for Natural Language Processing
  VentureBeat, Android Headlines, Gizbot, NextBigWhat
  
  
• Mentor for Google Launchpad Accelerator startups; Invited speaker for Launchpad Studio finance startups
  Google Blog, TechCrunch, VentureBeat
  
  
• Google ML Kit & Learn2Compress from Google AI
    Google AI blog: Custom On-Device ML Models with Learn2Compress
    TechCrunch, VentureBeat, ZDNet, Ars Technica: ML Kit powered by Learn2Compress makes it easy to add AI to mobile on Android and iOS
  
  
• On-Device Machine Intelligence with Neural Projections
  (video from ICML 2017)
  
  
• TensorFlow Lite and On-device Conversational Modeling
  (download from github)
  
  
• Hunt for the Zodiac TV Show Cast for HISTORY Channel's new five-part documentary series
  (Premiere: Tues Nov 14, 10/9C)
  
  
• On-device Machine Intelligence for smartwatches
  Google Research blog, VentureBeat, Engadget, ComputerWorld, Quartz
  
  
• Scalable Machine Learning Platform for NLP & computer vision that powers core Google products
  
  
• Photo Reply: Multimodal learning using computer vision and NLP
  announced at Google I/O 2016
  
  
• Allo (smart messaging powered by machine learning and NLP)
  announced at Google I/O 2016
  
  
• Smart Reply (automated email reply using machine learning)
  Google Research blog, Gmail blog, WIRED, TechCrunch, NYTimes, USA Today and other news media
  
  
• Cracking the code
  New Scientist magazine, ACM TechNews


Recent Talks


• Keynote talk
  Global Artificial Intelligence Conference, Santa Clara (virtual), October 2021


• Keynote talk
  AKBC Conference, October 2021


• Keynote talk on Powering Deep Learning with Structure
  Mexican NLP Summer School, NAACL, June 2021


• Invited talk on Building the Next-Generation AI: Small and Efficient Neural Computing
  Georgia Tech (Georgia Institute of Technology), Atlanta, February 2021


• Invited talk on Large-Scale Neural Graph Learning [video]
  Industry Advisory Board at University of California, Santa Cruz, California, January 2021


• Keynote talk on Neural Graph Computing at Scale
  TextGraphs at COLING, Barcelona, Spain, December 2020


• Invited talk on Efficient AI: Building Efficient Neural Computing Machines on the Edge & Cloud
  Toronto Machine Learning Summit, Toronto, November 2020


• Neural Structured Learning in TensorFlow [video]
  O'Reilly TensorFlow World Conference, California, October 2019


• Invited talk on Building Scalable & Privacy-Preserving Conversational AI Systems
  First Open Virtual Assistant Workshop, part of Stanford Human-Centered Artificial Intelligence (HAI) Conference,
  Stanford University, California, October 2019


• Invited talk on Building Neural Computing Machines at Scale in the Cloud and On-Device at the Edge
  Stanford University (CS Department), California, July 2019


• Invited talk on Building Neural Conversational Machines at Scale
  ReWork Deep Learning Summit, San Francisco, January 2019


• Keynote talk
  14TH International Workshop on Mining and Learning with Graphs at KDD, London, August 2018


• Keynote talk
  Deep Learning for Low-Resource NLP at ACL, Melbourne, July 2018


• Invited talk on Neural Structured Learning for Language and Vision
  SoCal NLP Symposium, Irvine, April 2018


• Learning On-Device Conversational Models
  MLPCD Workshop at NIPS, Long Beach, December 2017


• Structured Prediction and Neural Graph Learning at Scale
  RIKEN, Japan, October 2017


• Neural Graph Learning
  Deep Structured Prediction Workshop at ICML, Sydney, August 2017


• On-Device Machine Intelligence with Neural Projections
  TinyML Workshop at ICML, Sydney, August 2017 [video]


• Large-scale deep learning and graph methods for machine intelligence
  Association of Pattern Recognition and Image Analysis Summer School on Deep Learning, Spain, July 2017


• Keynote talk on Neural Graph Learning
  Workshop on Graph Algorithms and Machine Learning (GraML)
  IEEE Parallel and Distributed Processing Symposium, Orlando, June 2017


• Machine Intelligence at Massive Scale
  Big Data Innovation Summit, San Francisco, April 2017


• Invited Panelist for New York Times @ SXSW on The Promise (and Limits) of Artificial Intelligence
  Hosted by Gideon Lewis-Kraus, The New York Times Magazine, Austin, March 2017


Recent Publications [Google Scholar]


• On-Device Conversational Slot Extraction, SIDIAL 2021
• Efficiently Summarizing Text and Graph Encodings of Multi-Document Clusters, NAACL 2021
• Efficient Retrieval Optimized Multi-task Learning, arXiv 2021
• ProFormer: Towards On-Device LSH Projection Based Transformers, EACL 2021, arXiv 2020
• On-Device Text Representations Robust to Misspellings via Projections, EACL 2021
• Transductive Learning for Abstractive News Summarization, arXiv 2021
• Environment-agnostic Multitask Learning for Natural Language Grounded Navigation, ECCV 2020
• Low-Dimensional Hyperbolic Knowledge Graph Embeddings, ACL 2020
• GoEmotions: A Dataset of Fine-Grained Emotions, ACL 2020
• Ultra Fine-Grained Image Semantic Embedding, WSDM 2020
• ProSeqo: Projection Sequence Networks for On-Device Text Classification, EMNLP 2019
• PRADO: Projection Attention Networks for Document Classification On-Device, EMNLP 2019
• Graph Agreement Models for Semi-Supervised Learning, NeurIPS 2019
• On-device Structured and Context Partitioned Projection Networks, ACL 2019
• A2N: Attending to Neighbors for Knowledge Graph Inference, ACL 2019
• Deep Reinforcement Learning For Modeling Chit-Chat Dialog With Discrete Attributes, SIGDIAL 2019 (Best Paper Award)
• Efficient On-Device Models using Neural Projections, ICML 2019
• Transferable Neural Projection Representations, NAACL 2019
• Graph-RISE: Graph-Regularized Image Semantic Embedding, arXiv 2019
• GAP: Generalizable Approximate Graph Partitioning Framework, arXiv 2019
• Improved Graph based Semi-Supervised Learning, AISTATS 2019
• Conditional Utterance Generation with Discrete Dialog Attributes in Open-Domain Dialog Systems, Deep-Dial at AAAI 2019
• Fast & Small On-device Neural Networks for Short Text Natural Language Processing, On-Device ML (MLPCD 2) at NeurIPS 2018
• Modeling Non-Goal Oriented Dialog with Discrete Attributes, Conversational AI at NeurIPS 2018
• Self-Governing Neural Networks for On-Device Short Text Classification, EMNLP 2018
• PhotoReply: Automatically Suggesting Conversational Responses to Photos, The Web Conference (WWW) 2018
• Learning On-Device Conversational Models, MLPCD at NIPS 2017
• Neural Graph Machines, WSDM 2018
• ProjectionNet: Efficient On-Device Deep Networks, arXiv 2017
• Related Event Discovery, WSDM 2017
• Large Scale Distributed Semi-Supervised Learning Using Streaming Approximation, AISTATS 2016
• Hierarchical Label Propagation and Discovery for Machine Generated Email, WSDM 2016
• Smart Reply: Automated response suggestion for email, KDD 2016
• Semantic Video Trailers, MVRL at ICML 2016
• Conversational flow in Oxford-style debates, NAACL 2016


Chair & Organizer


• Area Chair AAAI 2021
• Area Chair (Knowledge Graphs) AACL 2020
• Area Chair (Machine Learning) EMNLP 2019
• Area Chair (Dialog and Interactive systems) NAACL 2019
• Co-Organizer ICML 2019 Joint Workshop on On-Device Machine Learning & Compact Deep Neural Network Representations (ODML-CDNNR)
• Co-Chair National Academy of Engineering (NAE) 2019 AI & Deep Learning track, German-American Frontiers of Engineering Symposium
• Co-Organizer NAACL 2019 3rd Workshop on Structured Prediction for NLP (SPNLP)
• Co-Organizer NeurIPS 2018 On-device Machine Learning Workshop (MLPCD2)
• Area Chair ACL 2017
• Area Chair COLING 2016
• Senior Program Committee IJCAI 2016
• Organizer Joint NAACL/ICML Symposium on NLP and Machine Learning 2013
• Workshop Chair NAACL HLT 2013
• Organizer First Workshop on Multilingual Modeling (ACL 2012)

Short Bio

Dr. Sujith Ravi is the Founder & CEO at Stealth AI/ML startup. Previously, he was the Director of Amazon Alexa AI where he led efforts to build the future of multimodal conversational AI experiences at scale. Prior to that, he was leading and managing multiple ML and NLP teams and efforts in Google AI. He founded and headed Google’s large-scale graph-based semi-supervised learning platform, deep learning platform for structured and unstructured data as well as on-device machine learning efforts for products used by billions of people in Search, Ads, Assistant, Gmail, Photos, Android, Cloud and YouTube. These technologies power conversational AI (e.g., Smart Reply), Web and Image Search; On-Device predictions in Android and Assistant; and ML platforms like Neural Structured Learning in TensorFlow, Learn2Compress as Google Cloud service, TensorFlow Lite for edge devices.

Dr. Ravi has authored over 100 scientific publications and patents in top-tier machine learning and natural language processing conferences. His work has been featured in press: Wired, Forbes, Forrester, New York Times, TechCrunch, VentureBeat, Engadget, New Scientist, among others, and also won the EACL Best Paper Award Honorable Mention in 2021, SIGDIAL Best Paper Award in 2019 and ACM SIGKDD Best Research Paper Award in 2014. For multiple years, he was a mentor for Google Launchpad startups. Dr. Ravi was the Co-Chair (AI and deep learning) for the 2019 National Academy of Engineering (NAE) Frontiers of Engineering symposium. He was also the Co-Chair for ACL 2021, EMNLP 2020, ICML 2019, NAACL 2019, and NeurIPS 2018 ML workshops and regularly serves as Senior/Area Chair and PC of top-tier machine learning and natural language processing conferences like NeurIPS, ICML, ACL, NAACL, AAAI, EMNLP, COLING, KDD, and WSDM.

Professional Activities (earlier version)

Chair


• Area Chair ACL 2017
• Area Chair COLING 2017
• Workshop Chair NAACL/HLT 2013


Organizer


• Joint NAACL/ICML Symposium on NLP and Machine Learning at NAACL / ICML 2013
• First Workshop on Multilingual Modeling at ACL 2012
• Natural Language Seminar series at USC Information Sciences Institute


Senior Program Committee Member


• IJCAI 2016
• CIKM 2015
• IJCAI 2011


Program Committee Member (recent, 2010 - 2017)


• NIPS
• ICML
• KDD
• AAAI
• WSDM
• ICWSM
• ACL
• EMNLP
• NAACL
• EACL
• IJCNLP
• BUCC: ACL Workshop on Building and Using Comparable Corpora
• IEKA: RANLP Workshop on Information Extraction and Knowledge Acquisition


Journal Reviewer


• IEEE TKDE - Transactions on Knowledge and Data Engineering (2012)
• Journal on Pattern Recognition (2011),
• ACM TIST - Transactions on Intelligent Systems and Technology Journal (2011),


Fun Fact


• I have an Erdos number = 3:
  
  • me => Andrei Broder => Bela Bollobas => Paul Erdos

Research

My main research interests lie in Artificial Intelligence areas, specifically machine learning, natural language processing (NLP), computer vision, multimodal learning, large-scale data mining, and other areas such as computational decipherment. Here are a few (selected) research topics that I have worked on.

• Semi-supervised and Unsupervised Learning for Structured Prediction
  • Semi-supervised learning for structured, exponentially large output spaces
  • Unsupervised learning via decipherment (ACL 2011)
  • Bayesian inference for cracking cryptographic ciphers (ACL 2011)
  • Unsupervised phonetic transliteration across languages (NAACL 2009)
  • Exact, approximate algorithms for model minimization; application to unsupervised part-of-speech tagging, supertagging, alignment (TACL 2014, ACL 2010, ACL 2009 [Nominated for Best Paper Award])


• Efficient Learning and Inference Algorithms
  • Scalable graph propagation for knowledge expansion
  • Reducing sampling complexity of topic models (KDD 2014 [Best Paper Award])
  • Fast clustering with exponential families (NIPS 2012)
  • Scalable unsupervised learning for machine translation (ACL 2013)
  • Parallel algorithms for unsupervised tagging (TACL/ACL 2014)
  • Summarization through submodularity and dispersion (ACL 2013)
  • Learning Probabilistic Finite State Machines (NAACL 2010)


• Multimodal Learning, Language Grounding and User Modeling in Social Media
  • Multimodal learning for inferring prototypical object names (AAAI 2015)
  • Understanding linguistic phenomena grounded in social content and user behavior (ICWSM 2014)
  • Studying predictability of language in social media (EMNLP 2012)


  Other Topics


• Data mining from the Web, Social Media
  • Graph Algorithms for Content Summarization
  • User modeling in Social Media
  • Computational Advertising


• Machine Translation and related problems
  • Machine Translation without parallel data
  • Machine Transliteration without parallel data
  • Unsupervised Word & Sub-word Alignment


• Parsing & Tagging
  • Unsupervised Part-of-Speech Tagging
  • Unsupervised Supertagging (using CCG)
  • Unsupervised Parsing
  • Parser Accuracy Prediction


• Computational Decipherment
  • Exact methods for cracking ciphers using minimal knowledge
  • Bayesian Decipherment for the famous Zodiac cipher, other homophonic ciphers


• Information Extraction and Discourse
  • Large-scale fact extraction from the Web
  • NLP/IR methods to aid users in finding relevant information from online discussion forums
  • Discourse analysis over discussion threads appearing in educational and collaborative settings


• Multilingual NLP
  • Italian Part-of-Speech Tagging, Supertagging
  • Spanish Temporal Expressions
  • Turkish Morphology Induction & Alignment

Publications