• Real-Time Open-Domain Question Answering (Paper Discussion)

    JJ Lake Business Center

    ** Please note that our location has changed Existing question answering (QA) models are not suitable for real-time usage because they need to process several long documents for every input query, which is computationally prohibitive. This paper introduces query-agnostic representations of phrases that can drastically speed up open-domain QA. In particular, their dense-sparse phrase encoding effectively captures syntactic, semantic, and lexical information of the phrases and eliminates the pipeline filtering of context documents. This model can be trained and deployed in a single 4-GPU server. The experiments on SQuAD-Open shows that this model is equal or better than previous models, with significantly less computational cost. Paper to read: Seo er al. "Real-Time Open-Domain Question Answering with Dense-Sparse Phrase Index." ACL 2019. https://arxiv.org/pdf/1906.05807 Presenter: Junling Hu This is part of the bi-weekly reading series. We come together to discuss cutting-edge AI topics and papers. One paper is selected as the major discussion topic. The meeting is led by one presenter, with group discussion and participation. Bring your questions and get answered. Socialize with other like-minded people. 6:30-7pm Meet and greet 7-8pm Paper presentation and group discussion 8-8:30 Additional social

    1
  • Machine Learning Security: Understanding Adversarial Examples

    Machine learning solutions are created at a rapid pace. Unfortunately, little focus is placed on security. Come to this talk to learn more about a specific category of attacks on machine learning called Adversarial Examples. Presenter: Abraham Kang Short biography of the speaker: Abraham Kang is CTO of GEEEE. He works on implementing Deepfakes detection and prevention utilizing machine learning. Kang has worked for various companies helping to drive AI, security, and development. He was Senior Director at Samsung Research America. He also worked as Principal Security Researcher for Fortify in their Software Security Research group. Prior to this, Abraham worked in application security for over 10 years. Kang has a B.S. from Cornell University and JD from Lincoln Law School of San Jose. Papers to read: 1. Carlini et al, “On Evaluating Adversarial Robustness”, Feb 2019 https://arxiv.org/pdf/1902.06705v2.pdf 2. Nicholas Carlini and David Wagner. “Adversarial examples are not easily detected: Bypassing ten detection methods”. In Proceedings of the 10th ACM Workshop on Artificial Intelligence and Security, 2017. https://arxiv.org/abs/1705.07263 Agenda: 6:30-7pm Meet and greet 7-8pm Lecture and Q&A 8-8:30 Additional social

    1
  • The Future of Dialogue Systems: A Summary of ACL 2019 Conference

    Dialog systems is the dominating theme at the just finished ACL 2019 Conference, July 28-August 2 in Florence, Italy. I will report on the papers, new ideas, and upcoming trend. Particularly I will report on the workshop “NLP for Conversational AI” and the major discussions from this workshop. After BERT and XLNet, deep learning has pushed its frontiers in NLP into question-answering, text summarization, story writing, and dialogue system. The most exciting one is dialogue systems, or chatbot. As many NLP researchers move into chatbot domain, we are seeing a breakthrough in this domain happening in the near future. Presenter: Junling Hu Papers to read: 1. ACL 2019 Conference program, with video link to each talk http://www.acl2019.org/EN/program.xhtml 2. ACL Workshop NLP for Conversational AI, with slides of each talk https://sites.google.com/view/nlp4convai/program This is part of weekly reading series. We come together to discuss the cutting-edge AI topics and papers. One paper is selected each week as the major discussion topic. The meeting is a mixture of presentation and white boarding, with group discussion and participation. Bring your questions and get answered. Socialize with other like-minded people. If you cannot attend in person, you can attend remotely by registering here https://zoom.us/webinar/register/WN_UgO3bl5FTiah5sCYBbkziA 6:30-7pm Meet and greet 7-8pm Group discussion 8-8:30 Additional social

    7
  • [Partner event] Big data and its impact on the planet

    This event is hosted by Analytics at Speed Palo Alto Meetup in partnership with IoT and Wearable Devices Meetup group. Talk content: Did you know scientists have predicted that unless radical improvements are made in the way we design computers, by 2040, computer chips will need more electricity than what our global energy production can deliver? Data is the new oil but using data to drive business outcomes is compute-intensive and unsustainable to the planet. Kx, the greenest and fastest database in the world, has a creative solution. Learn how we are enabling green analytics applications for our customers across the globe by processing data in real-time and in the most energy-efficient way. Our esteemed panel members will discuss the relevance of green analytics for next-generation data application. Agenda: 6:30 – 7:00 pm – Registration and Networking 7:00 – 7:10 pm. – Welcome and Introduction to Kx 7:10 – 7:45 pm – Panel Discussion on Green Analytics 7:45 – 8:15pm – Kx Solution Overview & Demo 8:15 – 9:00 pm – Light bites, and networking Panel Members: 1. Sri Satish Ambati, CEO H20.ai 2.Gayathri Gopalakrishnan, CEO, Ecoformatics 3. Michael Falcon, Investment Director, LG Technology Ventures 4. Christopher Conklin, Chief Engineering Officer, Carfit 5. Jimmy Moore, CEO, Untangle AI 6. Uday Ayyagari, Lead Architect, Crowdz In attendance: The Aston Martin Red Bull Racing static showcar We look forward to meeting all of you! You can also check out our partner's page: https://www.meetup.com/Silicon-Valley-IoT-and-Wearable-Devices/events/263174429/ Our esteemed panellist: Gayathri Gopalakrishnan is an award-winning data scientist and environmental engineer who has been pioneering the integration of clean technology and data science since 2003. She worked as an environmental scientist in academia and national laboratories, led Facebook's mapping and AI initiative in their sustainability team and as an engineer and data scientist in several Silicon Valley startups. Most recently, she is the founder and CEO of Ecoformatics, an education startup teaching data science skills to clean technology professionals. Michael Falcon has over twenty years of venture capital and operating experience, with an emphasis on Deep Tech and information technology industry sectors. Prior to LG Technology Ventures, Michael was a Managing Partner of Bandgap Ventures, which focuses on physical science technology investing in the areas of advanced materials, manufacturing, and electronics. Michael began his work in venture capital at Alara Capital (formerly Guggenheim Venture Partners), which was the venture capital group of Guggenheim Partners, a global financial services firm with more than $200 billion in assets under management.

  • [Partner event] Venture Capital Panel: Investment and Innovations in AI Startups

    The deadline to register and pay online is 2 pm the day of the event. Park in student lots 8, 9, 6, or 2. Parking is okay in these lots and free for this event. (Normal campus parking regulations won't be enforced.) Lot 8 is closest to the meeting space. Go to College Center - Building 10, Bay View Dining Room (2nd floor). Please review the campus map in advance so you can find the meeting space and parking seamlessly. http://collegeofsanmateo.edu/map/docs/CSMCampusMap.pdf PLEASE READ THE ENTIRE EVENT DESCRIPTION. There is a cover charge for this event. A "YES" RSVP on our meetup site does not confirm your reservation. Space is limited. Please only RSVP "Yes" if you will be attending. Take advantage of early responder pricing! Slots are limited. First come, first served! First Responders $12 plus Eventbrite fees Second Responders $15 plus Eventbrite fees Third Responders $17 plus Eventbrite fees Fourth Responders $20 plus Eventbrite fees In order to attend, you must register and pay here: https://www.eventbrite.com/e/venture-capital-panel-investment-and-latest-innovations-in-ai-startups-tickets-64640687035 Please bring your printed ticket to the event. Or, pay $25 (cash) at the door. A distinguished panel of venture capitalists will discuss investment AI startups and the latest innovations in AI -- and what to expect throughout 2019. (This is not a pitch event, so no startups will be pitching to the panel.) If you are an AI startup and would like space on a demo table for the networking portion of the event, please send an email message to: [masked] Put this in the subject header: T 7/23 Startup Demo Table Agenda: 6:00 pm to 7:00 pm Check In, Food, Networking Reception 7:00 pm to 8:30 pm Panel Discussion, Q & A 8:30 pm to 9:00 pm More Networking About the Panelists: Jennifer Vancini is a general partner with Mighty Capital. Jennifer has over 20 years of experience across a broad range of technologies, with extensive experience in the mobile and security industries. She has held management and executive roles in start-ups and innovation units of large organizations including Price Waterhouse, Nokia, the Symbian Foundation and Telefonica, and has served on several boards. Nuno Goncalves Pedro is a venture partner with Grishin Robotics, a Sand Hill Road venture capital firm focused on smart hardware investments. He is also the founder and managing partner of Strive Capital, a quantitative early stage venture capital fund focused on early-stage consumer mobile and deep software (artificial intelligence and blockchain). Homan Yuen is a Partner at Fusion Fund, an early-stage venture capital firm focused on being the first institutional capital into innovative companies bringing technologies to market. The firm has over 50 industrial, enterprise, and healthcare technology companies in the portfolio. Homan received his Ph.D. and M.S. in Electrical Engineering and Materials Science from Stanford University and his B.A. in Physics from UC Berkeley. Homan received his Ph.D. and M.S. in Electrical Engineering and Materials Science from Stanford University and his B.A. in Physics from UC Berkeley. George Arabian is a start up veteran and managing partner of Steelhead Ventures, a micro venture capital fund with seed stage investments in companies his brother and partner, Gary and him truly believe in. He works with founders, other venture partners, thought leaders and executive teams who are making a positive difference in their communities and the future of our world. Steelhead’s portfolio includes 30+ visionary founders. About the Moderator: Roger Royse is the founder of Royse Law Firm, PC, a business and tax law firm with offices in Northern and Southern California (www.rroyselaw.com (http://www.rroyselaw.com/)). Roger practices in the areas of corporate and securities law, domestic and international tax, mergers and acquisitions, and fund formation.

  • Multiagent Deep Reinforcement Learning: Coordination through Social Influence

    In 2019, we are seeing the rise of multiagent systems: From AlphaStar winning multi-player games with Nash strategy to research on self-driving cars interacting with other self-driving cars. There is the coming reality that multiple intelligent agents co-exist in our world. This paper from DeepMind proposed a unified mechanism for achieving coordination and communication in Multiagent Reinforcement Learning, through rewarding agents for having causal influence over other agents’ actions. Agents interact in decentralized manner where coordination can only be done through organic communication. This paper was presented in ICML 2019. Papers to read: Jaques, Natasha, Angeliki Lazaridou, Edward Hughes, Caglar Gulcehre, Pedro Ortega, Dj Strouse, Joel Z. Leibo, and Nando De Freitas. "Social Influence as Intrinsic Motivation for Multi-Agent Deep Reinforcement Learning." In ICML 2019 http://proceedings.mlr.press/v97/jaques19a/jaques19a.pdf Background paper: Leibo, J. Z., Zambaldi,V., Lanctot, M., Marecki, J., and Graepel, T. "Multi-agent reinforcement learning in sequential social dilemmas". In Proceedings of the 16th Conference on Autonomous Agents and MultiAgent Systems, 2017. https://arxiv.org/pdf/1702.03037.pdf This is part of weekly reading series. We come together to discuss the cutting-edge AI topics and papers. One paper is selected each week as the major discussion topic. The meeting is a mixture of presentation and white boarding, with group discussion. Bring your questions and get answered. Socialize with other like-minded people. You can attend in person or remotely. If you attend remotely, register here: https://zoom.us/webinar/register/WN_GymNA3UBRj6RTPFxH3x4JA You will receive a link to join the live presentation 6:30-7pm Meet and greet 7-8pm Group discussion 8-8:30 Additional social

    3
  • XLNet: A New NLP Method that outperforms BERT

    385 Moffett Park Dr

    XLNet is a new method for NLP from Google Brain that was released on June 19, 2019. XLNet significantly improves upon BERT on 20 tasks and achieves state-of-the-art results on 18 tasks including question answering, natural language inference, sentiment analysis, and document ranking. For example, in SQuAD 2.0, XLNet achieved 86% in exact match, while BERT only reached 79%. We will review XLNet and its difference from BERT. Papers to read: Yang et al. "XLNet: Generalized Autoregressive Pretraining for Language Understanding." arXiv:[masked] (June 2019). https://arxiv.org/abs/1906.08237 Background paper: Dai et al. "Transformer-xl: Attentive language models beyond a fixed-length context." ACL 2019. https://arxiv.org/pdf/1901.02860 Github (code + pretrained models): https://github.com/zihangdai/xlnet You can attend in person or remotely. If you attend remotely, register here: https://zoom.us/webinar/register/WN_j29BAwTlTOmGl8ReJy2dmA You will receive a link to join the live presentation This is part of weekly reading series. We come together to discuss the cutting-edge AI topics and papers. One paper is selected each week as the major discussion topic. The meeting is a mixture of presentation and white boarding, with group discussion and participation. Bring your questions and get answered. Socialize with other like-minded people. 6:30-7pm Meet and greet 7-8pm Group discussion 8-8:30 Additional social

    1
  • Dynamic Medical Measurement Using Deep Reinforcement Learning

    If a patient is in critical condition, what and when should be measured to forecast detrimental events, especially under the budget constraints? This paper applies deep reinforcement learning (RL) to jointly minimizes the measurement cost and maximizes predictive gain, by scheduling dynamic measurements. The result was tested in a real-world ICU mortality prediction task, and reduced the total number of measurements by 31% or improve predictive gain by a factor of 3. This paper was presented in ICML 2019, and the authors are from University of Toronto and Vector Institute (An AI research institute led by Geoff Hinton). Paper to read: Chun-Hao Chang, M. Mai and A. Goldenberg, "Dynamic Measurement Scheduling for Event Forecasting Using Deep RL", ICML 2019 https://arxiv.org/pdf/1901.09699 Data and code: https://github.com/zzzace2000/autodiagnosis This is part of weekly reading series. We come together to discuss the cutting-edge AI topics and papers. One paper is selected each week as the major discussion topic. You can attend in person or remotely. If you attend remotely, register here https://zoom.us/webinar/register/WN_lqDHHQQHTYK_QGPU3OJt5g After registration, you will receive a link to join the presentation. 6:30-7pm Meet and greet 7-8pm Group discussion 8-8:30 Additional social

    6
  • Reinforcement Learning for Real Life

    385 Moffett Park Dr

    Talk abstract: Reinforcement Learning (RL) has achieved success in many domains such as AlphaGo, Starcraft game and robotics. What are the other real life applications? What are the challenges and opportunities? In this talk, I will summarize the talks and themes of the ICML 2019 workshop Reinforcement Learning for Real Life that I helped to organize. Our workshop brought together researchers and practitioners from industry and academia on applying RL to real life scenarios. This workshop was one of the most attended at ICML, with more than 550 people filled the room. It was also highlighted in Venture Beat's June 14 report (https://venturebeat.com/2019/06/14/ai-weekly-icml-2019-top-papers-and-highlights/). After a summary of the workshop, I will provide my viewpoint and prediction on future trend of RL applications. Workshop link: https://sites.google.com/view/RL4RealLife Speaker Bio: Yuxi Li is the founder of attain.ai. He has more than ten years experience in reinforcement learning, machine learning, and AI. He was a Co-Chair of ICML 2019 Reinforcement Learning for Real Life Workshop. He published the influential article Deep Reinforcement Learning: An Overview on arXiv, which was highly cited. He is the author of an upcoming book on Deep Reinforcement Learning. He was a Program Committee Member of AAAI 2019. He was also a co-organizer for AI Frontiers Conference (aifrontiers.com) in Silicon Valley in 2017 and 2018. Yuxi received his PhD in computer science from the University of Alberta. **This lecture will be delivered on screen. We will watch and discuss the webinar together. You can also join remotely (See registration link below) If you join remotely, register here (Limited slots available) https://zoom.us/webinar/register/WN_XW-PDSWhQdqToay9YERiWA You will receive a confirmation email with link to join the webinar. If you come to the meetup, no need to register as we will watch together. Meetup agenda: 6:30-7pm Meet and greet 7-8pm Webinar 8-8:30 Additional social

  • ICML 2019 paper highlight

    385 Moffett Park Dr

    ICML (International Conference of Machine Learning) is one of the largest machine learning conferences. It features cutting-edge research on deep learning, reinforcement learning, and other machine learning areas. The 2019 conference is held June 10 -June 15 in Long Beach, California. This year's conference has accepted total 774 papers (out of 3424 submissions, with 22.6% acceptance rate). We will review the highlight of these papers, and understand the hot research areas and trends in 2019. We will discuss the following questions: 1. What are the major themes in deep learning this year? 3. What are the major themes in reinforcement learning this year? 4. Other trends to watch out Papers to read: ICML 2019 accepted Papers https://icml.cc/Conferences/2019/Schedule?type=Oral Papers in different tracks: https://icml.cc/Conferences/2019/ScheduleMultitrack?text=&session=&day=2019-06-11&event_type= This is part of weekly reading series. We come together to discuss the cutting-edge AI topics and papers. One paper is selected each week as the major discussion topic. The meeting is a mixture of presentation and white boarding, with group discussion and participation. Bring your questions and get answered. Socialize with other like-minded people. 6:30-7pm Meet and greet 7-8pm Group discussion 8-8:30 Additional social

    4