风险投资人眼中的人工智能.pdfVIP

Re.Work Investing in Deep Learning 1st December 2015 Nathan Benaich, Ph.D Playfair Capital Talk outline Part 1 Why now, market entry, challenges investors/operators consider, product dev Part 2 Investments and opportunities About me/Playfair MPhil/PhD in experimental/computational cancer research. 2+ year tech VC. London-based investment firm focusing on early stage technology companies that change the way we live and work. Team of 5. Young, ambitious, founders first. Our remit is global and our focus is on core technologies and user experiences. 25% of portfolio is AI: speech/text, NLP, vision, deep learning We want to take risk on technologists creating new markets or reinventing existing ones. Why now AI is one of the most exciting and transformative opportunities of our time. 2 billion smartphones and users addicted, 40% world online → data assets Plummeting cost of compute and storage, compute power → pace of innovation accelerating. AI already out in the wild improving performance (search, music, ads, fintech) Improved learning methods, architectures and software infrastructure Build systems that learn about the world to make predictions and improve with use Market entry Connect data silos to facilitate information retrieval/insights: DueDil, Premise Go deep on a specific application: e.g. Ravelin with fraud detection Provide a general service: plug in a data and return predictions: H20, Seldon Optimise workflows and repetitive processes: e.g. Gluru, SwiftKey Autonomous systems for physical world: e.g. Tesla Core long-term RD: DNN Research, DeepMind Key consideration: open source and cheap service providers are eroding technology barriers. Data and talent move the needle. People, products and data. Challenges investors focus on - Operational - Long-term RD or short term monetisation? Performance = acceptable. - Sha