AI By the Bay 2017
| title | author | length |
|---|---|---|
| AI and Self-Driving Cars | Danny Shapiro | 42m |
NVIDIA argues deep learning replaces hand‑coded perception by training models on real and simulated data and deploying them on an in‑car supercomputer. A… | ||
| State of AI: Chatbots, Robots & Virtual Reality, Oh My! | Michael Ludden | 33m |
Overview of concrete offerings and demos: IBM exposes ~30 REST API services on Bluemix with a 30-day free trial and a persistent free tier for prototyping… | ||
| Developing and Deploying Models at Scale — without Spark! | Eduardo Ariño de la Rubia | 27m |
Live demo trains an h2o deep learning model on the Airline dataset (about 10 million rows) on a single X1 instance using 128 cores and terabytes of RAM, then… | ||
| The Real Future of AI Panel | Alexy Khrabrov, Lukas Biewald, Lyle Ungar, Peter Norvig, Shubha Nabar, SriSatish Ambati, Stuart Russell | 57m |
This talk video is a node in devreal.ai, the developer community graph.
It connects to:
Talk: The Real Future of AI Panel —… | ||
| Bots Panel | Andrew Burgert, David Hall, Dennis Yang, Pete Skomoroch, Sonal Gupta, Yishay Carmiel | 37m |
Panelists argue bots already process massive live data and show early wins — 1.2 billion messages processed — but true, flexible conversation that handles… | ||
| AI VC Panel | Adam Kell, Alexy Khrabrov, Benjamin Levy, Bradford Cross, David Beyer, Jake Flomenberg | 39m |
This talk video is a node in devreal.ai, the developer community graph.
It connects to:
Talk: AI VC Panel —… | ||
| Panel: Smart Cars and Smart Cities | 57m | |
This talk video is a node in devreal.ai, the developer community graph.
It connects to:
Talk: Panel: Smart Cars and Smart Cities —… | ||
| AI: from an Idea to the Customer Panel | Alexy Khrabrov | 54m |
This talk video is a node in devreal.ai, the developer community graph.
It connects to:
Talk: AI: from an Idea to the Customer Panel —… | ||
| Interview | Dennis Yang | 04m |
Argues that conversation becomes the primary interface to information and that conversational analytics are the toolchain to get there: Dashbot collects… | ||
| Interview | Lukas Biewald | 04m |
Argues that practical AI succeeds when massive labeled examples exist — cites billions of selfies powering Facebook face recognition and parallel corpora… | ||
| Interview | Andrew Burgert | 03m |
Claims a practical stack for conversational UIs that uses machine learning to interpret user intent and pull entities from utterances, improving handoffs to… | ||
| Interview | Jeremy Howard | 03m |
Argues that deep learning does not require lots of data, many GPUs, or years of math education and shows practical steps to make it accessible. Contrasts… | ||
| Interview | Bradford Cross | 04m |
Claims that large-data prediction and ranking problems are largely solved by current machine learning stacks and cites machine learning subfields like natural… | ||
| Interview | Tom Brown | 05m |
Defines intelligence = ability to achieve goals efficiently across varied circumstances and calls artificial simply nonbiological, claims supervised learning… | ||
| Interview | Adam Kell | 03m |
AI is defined here as a system that reinforces itself via a self-containing feedback loop rather than explicit code, and practical progress comes from… | ||
| Interview | Benjamin Levy | 04m |
Argues investors should back narrow AI that uses proprietary data to solve industry-specific problems, not chase general intelligence; credits recent gains to… | ||
| Interview | Francois Chollet | 04m |
Argues that current wins are concrete pattern‑matching systems — computer vision for object recognition and speech recognition — enabled by deep learning, but… | ||
| BachBot: Composing Bach Chorales using Deep Learning | Feynman Liang | 26m |
BachBot uses a three-layer LSTM trained on the Music21 corpus of all known Bach chorales after transposing to C major/minor, quantizing durations, and encoding… | ||
| AI at Stitch Fix | Chris Moody | 26m |
Shows how interpretability comes from concrete choices: replacing point embeddings with distributions so vectors carry uncertainty ellipses, using SVD for word… | ||
| The Advantages of Deep Learning with Recurrent Neural Networks | Mike Tamir | 31m |
Realtime anomaly detection using LSTM RNNs frames forecasting as predict-and-analyze residuals where each metric’s next-step forecast is compared to the… | ||
| Deploying and Scaling Spark ML and Tensorflow AI Models | Chris Fregly | 33m |
Shows a working pipeline that trains, packages and serves models with Tensorflow, Spark ML, Jupyter, Docker and Kubernetes plus NetflixOSS microservices, with… | ||
| Deep Learning is Like Water (it's Everywhere!) | Arno Candel | 42m |
Demonstration of a distributed H2O.ai stack that parses 116 million rows in 8 seconds across a 10-node cluster (≈320 cores), compresses 12GB to ~2GB, trains… | ||
| Attention and Memory in Deep Learning Networks | Stephen Merity | 42m |
Shows that attention removes the encoder bottleneck in sequence‑to‑sequence translation so sentence length (a 38‑word example) stops collapsing accuracy after… | ||
| Deep Learning with the GPUs in Production | Adam Gibson | 42m |
This talk video is a node in devreal.ai, the developer community graph.
It connects to:
Talk: Deep Learning with the GPUs in Productio... —… | ||
| Machine Learning, The Right Way | Vitaly Gordon | 37m |
Argues that labeled data is the bottleneck not model hype, shows simple logistic regression experiments where a single C hyperparameter shifts accuracy from… | ||
| Host Sponsor Welcome | Joel Horowitz | 10m |
IBM positions itself as enabling cognitive computing at global scale, offering public access to quantum machines via IBM Bluemix quantum access and promoting… | ||
| Interview | Stuart Russell | 06m |
Survey of AI that contrasts perceptual gains from deep learning with long-range decision work, citing AlphaGo as offline machine learning plus online search… | ||
| Interview | Ted Graham | 04m |
Works on building autonomous and connected cars at General Motors and describes concrete uses: mapping for localization, ML models for passenger safety, sales… | ||
| Interview | Mark Harris | 04m |
Defines AI as software that actually understands rather than just acts, points to improving perception that can recognize images faces and dangerous… | ||
| Interview | Elliot Katz | 04m |
Claims that autonomous vehicle tech reaches 98% complete for capability, with the last 2% hard to solve; uses deep learning as the mechanism to handle rare… | ||
| Interview | Xin Wei Ngiam | 03m |
Defines AI as machines that learn from data to refine decisions and claims it fits automation and optimization at big scale, especially real-time use cases… | ||
| Interview | Doug Davenport | 03m |
Argues that AI will be a force multiplier for transportation and transit, claiming it can convert antiquated traffic and transit systems into interconnected… | ||
| Interview | Michael Ludden | 06m |
Defines AI as essentially modern machine learning algorithms and says current wins come when systems auto-correct from a large corpus of data to provide… | ||
| Interview | Christopher Heiser | 03m |
Renovo describes a working platform that ingests gigabytes of data per hour per car and runs internal machine learning systems for anomaly detection… | ||
| Interview | David Silver | 04m |
Artificial intelligence shifts self-driving systems from handcoded rules developed over the last 20 years toward deep learning trained on large diverse… | ||
| Measuring Psychological Traits using Social Media | Lyle Ungar | 38m |
This talk video is a node in devreal.ai, the developer community graph.
It connects to:
Talk: Measuring Psychological Traits using Soc... —… | ||
| Tackling the Limits of Deep Learning | Richard Socher | 39m |
Argues for one joint QA framework and shows concrete modules that work: a dynamic memory network with an episodic memory module that makes multiple attention… | ||
| Measuring AI Capabilities with OpenAI Gym & Universe | Tom Brown | 37m |
OpenAI releases a tiny API—env.make, step, reset—and a large catalog of environments so researchers can measure concrete capability gains. OpenAI Gym supplies… | ||
| Computer Vision for Coders | Jeremy Howard | 37m |
This talk video is a node in devreal.ai, the developer community graph.
It connects to:
Talk: Computer Vision for Coders —… | ||
| Machine Learning Startups | Bradford Cross | 44m |
Four startups since 2009 and a VC fund founded in 2011 lead to one clear prescription: build full‑stack, domain‑specific products with proprietary data, not… | ||
| Advances in Deep Learning for Mathematical Theorem Proving | Francois Chollet | 40m |
Google project augments a brute‑force ATP called improver with deep learning to rank axioms and guide proof search, using the Mizar (58k statements) corpus. A… | ||
| The Future of (Artificial) Intelligence | Stuart Russell | 46m |
Shows a probabilistic programming system called Bayesian Logic (BLOG) that generates large, dynamic graphical models and runs MCMC over data from ~150 seismic… | ||
| Building Training Data for Autonomous Driving | Darren Nakhuda | 17m |
Company pitches training data as a service and reports $27M raised while explaining concrete tooling choices: initial bounding-box and full-image segmentation… | ||
| ↑↑↓↓← → ← → Autonomous | Stefan Seltz-Axmacher, Kartik Tiwari | 16m |
Starsky Robotics already moves freight on highways using remote control to/from distribution centers and autonomous lane-keeping plus adaptive cruise on the… | ||
| Interview | John Eggert | 02m |
Velodyne describes how customers apply AI for modular perception to extract lanes, obstacles and object classes from high definition 3D LiDAR rather than hard… | ||
| Interview | Josh Hartung | 04m |
Company builds infrastructure and safety frameworks for autonomous systems, arguing AI as generalized computing engine (cites an Nvidia framing) and showing AI… | ||
| Interview | William Baumgardner | 02m |
Argues that combining autonomous vehicles with big data can fundamentally reshape urban mobility, changing how people move, access services and how cities… | ||
| Interview | Wolfgang Juchmann | 05m |
NVIDIA uses a provided car platform and camera feed to train a neural network that learns lane keeping and even infers turn intent when the indicator is on, so… | ||
| Interview | George Hotz | 02m |
Argues that real progress splits into learning world models separate from policy and tackling long-horizon reinforcement learning where sparse rewards force… | ||
| Interview | Jur van den Berg | 04m |
Argues that modern deep learning already trains networks equal to humans at detecting cars and that researchers still do not know why those models work so… | ||
| Interview | Danny Shapiro | 04m |
Company builds a hardware and software stack using dr DX and drive work software to run deep learning on GPUs for perception and end-to-end driving. Claims… | ||
| Interview | Gabor Melli | 03m |
Defines AI as machines showing adaptive behavior and cites a Sony product that models user behaviors to personalize experience, reports concrete wins using… | ||
| Interview | Leon Katsnelson | 05m |
Argues that AI is augmented intelligence that explains why only 13 companies beat the S&P 500 consistently over five years, credits those winners for turning… | ||
| Interview | Daniel Golden | 11m |
A cloud post-processing suite uses deep learning-based contouring to auto-segment cardiac MRI ventricles, trained on thousands of clinically annotated studies… | ||
| Interview | Feynman Liang | 03m |
Describes a production AI stack that automatically matches workers to gigs, prices work and predicts escalations by monitoring development lifecycle metrics… | ||
| Interview | Chris Moody | 05m |
Argues that useful AI is an instrument that helps people do science better, not a closed oracle, and shows concrete concerns: companies hoard training sets so… | ||
| Interview | Eduardo Ariño de la Rubia | 03m |
Argues that AI is the next step beyond current machine learning and that Domino Data Lab platform is what customers use to train those models, enabling data… | ||
| Interview | Chris Fregly | 05m |
Focuses on the gritty post-training work: how TensorFlow vs scikit-learn require different deployment paths, efforts to optimize models after training, and… | ||
| Interview | Arno Candel | 04m |
Claims that machine learning is already replacing hand‑coded decision rules by producing billion auto-generated rules that outperform simple if‑then logic, and… | ||
| Interview | Stephen Merity | 02m |
Argues that real progress comes from concrete model abilities not broad "AI" labels and demonstrates where deep learning wins on visual feature extraction but… | ||
| Mike Tamir Interview | Mike Tamir | 04m |
Argues that recent progress ties specific deep learning methods to capabilities: convolutional neural nets drive image recognition, 1D convolutions push… | ||
| Adam Gibson Interview | Adam Gibson | 03m |
AI splits into subfields with production work defaulting to machine learning and deep learning; Skymind focuses on deploying models for Fortune 2000 customers… | ||
| Vitaly Gordon Interview | Vitaly Gordon | 04m |
Argues that consumer AI is mature but enterprise and everyday services lag, and describes building a platform that moves models from demos into running… | ||
| Autonomy at Scale: The Need for Interop | Christopher Heiser | 32m |
Renovo describes an automotive-grade platform aiming for SAE Level 4–5 geofenced electric fleets with no steering wheel, validated on a skid‑pad DeLorean… | ||
| Becoming a Self-Driving Car Engineer | David Silver | 31m |
Udacity teaches a nine-month nanodegree that covers deep learning, computer vision, sensor fusion, localization, control and path planning with industry… | ||
| The Role of High Definition 3D LiDAR in Self Driving Cars | John Eggert | 28m |
Velodyne describes building rotating multi-channel time-of-flight LiDAR that produces instantaneous 3D point clouds—400,000 pulses per blink enabling real‑time… | ||
| Bringing Automated Driving Software to the Production Vehicle | Josh Hartung | 40m |
This talk video is a node in devreal.ai, the developer community graph.
It connects to:
Talk: Bringing Automated Driving Software to t... —… | ||
| Autonomous Vehicles 101 | Wolfgang Juchmann | 40m |
This talk video is a node in devreal.ai, the developer community graph.
It connects to:
Talk: Autonomous Vehicles 101 —… | ||
| Self-Driving Lessons from Comma AI | George Hotz | 41m |
An open source level‑2 driving agent openpilot runs today on a <$1,000 Comma Neo device and claims ~10 miles between disengagements now with software updates… | ||
| Technical Challenges of Self-Driving Trucks | Jur Van Der Berg | 34m |
This talk video is a node in devreal.ai, the developer community graph.
It connects to:
Talk: Technical Challenges of Self-Driving Tru... —… | ||
| nothing matches | ||