MLconf SF 2017
2017-11-10
San Francisco, United States
[MLconf SF 2014, MLconf SF 2015, MLconf SF 2016, MLconf SF 2017, MLconf SF 2019]
who it's for
platform engineers, data scientists, research scientists, engineering managers, applied ML researchers, product managers
topics
MXNet hierarchical parameter server; Gluon hybridization; tensor contraction low-rank regression; Tensorly with MXNet backend; GADF image time series; Inception v3 embeddings; sketch-rnn variational autoencoder; Disklavier music dataset modeling; Kubernetes orchestration for ML; Evolution Strategies for training; concept network reinforcement learning; skip-gram style TinVec embeddings
| title | author | length |
|---|---|---|
| Anima Anadkumar, Principal Scientist, Amazon Web Services, Endowed Professor, CalTech | Anima Anadkumar | 42m |
MXNet demonstrates a hierarchical parameter server that places a GPU as the level‑one server to cut GPU‑GPU vs CPU‑GPU bandwidth costs and records 91%… | ||
| Ashfaq Munshi, ML7 Fellow, Pepperdata | Ashfaq Munshi | 24m |
Pepperdata ingests ~300 metrics per task every 5s (one 570-node customer yields ~41M points/min; company-wide ~300 trillion points) and maps each series to… | ||
| Daniel Shank, Data Scientist, Talla | Daniel Shank | 17m |
Goal is to detect repeated help questions and auto-serve past answers by matching new queries to historical chat lines. Uses active learning starting from… | ||
| Doug Eck, Research Scientist, Google | Doug Eck | 31m |
Magenta experiments ship concrete code and demos: QuickDraw (~1B sketches) feeds a sketch-rnn variational autoencoder that models strokes as Δx, Δy plus… | ||
| Jonas Schneider, Head of Engineering for Robotics, OpenAI | Jonas Schneider | 19m |
OpenAI builds end-to-end ML infrastructure that runs on Kubernetes across Azure/AWS and private datacenters, mixing TensorFlow with custom ops, MPI for sync… | ||
| Josh Wills, Head of Data Engineering, Slack | Josh Wills | 24m |
Model building time collapses from months to days using XGBoost, embeddings, and emerging neural architecture search tools, enabling seismic oil imaging that… | ||
| Dr. June Andrews, Principal Data Scientist, Wise.io, From GE Digital | June Andrews | 20m |
Claims a deployed real-time production anomaly detector that ingests aircraft engine telemetry, flags out-of-normal conditions, and surfaces a… | ||
| LN Renganarayana, Architect, ML Platform and Services, Workday San Francisco | LN Renganarayana | 22m |
Workday builds enterprise ML products that initially overpromises—an early push for 40 apps yields one usable MVP—because k‑fold shows 95% accuracy but… | ||
| Matineh Shaker, Artificial Intelligence Scientist, Bonsai | Matineh Shaker | 17m |
Concept Network Reinforcement Learning decomposes hard RL tasks into shallow trees of reusable sub-concepts (policies) and selector nodes, and shows industrial… | ||
| Michael Alcorn, Sr. Software Engineer, Red Hat Inc. | Michael Alcorn | 19m |
Shows concrete systems: trains word2vec on click streams as URL2vec (treating each day as a sentence) to surface contextually similar docs that TF‑IDF misses… | ||
| Rushin Shah, Engineering Manager, Facebook | Rushin Shah | 19m |
Facebook deploys a production NLP stack that labels tens of thousands of topics, runs sequence labeling with LSTM/CRF and CNN pipelines, and uses a DeepText… | ||
| Dr. Steve Liu, Chief Scientist, Tinder | Steve Liu | 25m |
Tinder builds a production recommender called TinVec that embeds users from swipes-only data using a skip-gram-style embedding (feedforward NN with hidden… | ||
| Ted Willke, Sr. Principal Engineer, Intel | Ted Willke | 28m |
Uses low-flying snotbot drones and shore photos to do two practical ML systems on Alaska 2017 expeditions: a fluke-ID pipeline that extracts trailing-edge… | ||
| Xavier Amatriain, Cofounder & CTO, Curai | Xavier Amatriain | 27m |
Healthcare systems still give clinicians 15 minutes per visit, and medical errors cause about 400,000 deaths annually, so the approach combines ML models… | ||
| nothing matches | ||