Schedule: Models and Methods sessions: Artificial Intelligence & Machine Learning Conference: O'Reilly AI

9:00am–12:30pm Tuesday, September 10, 2019

Location: LL21 E/F

Using Keras to classify text with LSTMs and other ML techniques

Secondary topics: Deep Learning, Deep Learning tools, Machine Learning, Text, Language, and Speech

Lukas Biewald (Weights & Biases)

Average rating:

(4.25, 4 ratings)

Join Lukas Biewald to build and deploy long short-term memories (LSTMs), grated recurrent units (GRUs), and other text classification techniques using Keras and scikit-learn. Read more.

9:00am–12:30pm Tuesday, September 10, 2019

Location: 231

Getting started with Kubeflow

Secondary topics: Deep Learning tools, Machine Learning

Skyler Thomas (MapR)

Average rating:

(4.25, 4 ratings)

The popular open source Kubeflow project is one of the best ways to start doing machine learning and AI on top of Kubernetes. However, Kubeflow is a huge project with dozens of large complex components. Skyler Thomas dives into the Kubeflow components and how they interact with Kubernetes. He explores the machine learning lifecycle from model training to model serving. Read more.

1:30pm–5:00pm Tuesday, September 10, 2019

Location: LL21 A/B

Deep learning at scale: Tools and solutions

Secondary topics: Deep Learning, Deep Learning tools, Hardware, Machine Learning

Angela Wu (Determined AI), Sidney Wijngaarde (Determined AI), Shiyuan Zhu (Determined AI), Vishnu Mohan (Determined AI)

Average rating:

(4.67, 3 ratings)

Success with DL requires more than just TensorFlow or PyTorch. Angela Wu, Sidney Wijngaarde, Shiyuan Zhu, and Vishnu Mohan detail practical problems faced by practitioners and the software tools and techniques you'll need to address the problems, including data prep, GPU scheduling, hyperparameter tuning, distributed training, metrics management, deployment, mobile and edge optimization, and more. Read more.

11:55am–12:35pm Wednesday, September 11, 2019

Location: LL21 C/D

Document understanding: Extracting structured information from financial images and forms

Secondary topics: Computer Vision, Data, Data Networks, Data Quality, Machine Learning, Text, Language, and Speech

Joy Rimchala (Intuit), TJ Torres (Intuit), Xiao Xiao (Intuit), Hui Wang (Intuit)

Average rating:

(5.00, 1 rating)

Document understanding is a company-wide initiative at Intuit that aims to make data preparation and entry obsolete through the application of computer vision and machine learning. A team of data scientists, Joy Rimchala, TJ Torres, Xiao Xiao, and Hui Wang, detail the design and modeling methodologies used to build this platform as a service. Read more.

1:45pm–2:25pm Wednesday, September 11, 2019

Location: LL21 C/D

Sequence to sequence modeling for time series forecasting

Secondary topics: Deep Learning, Machine Learning, Temporal data and time-series

Arun Kejariwal (Independent), Ira Cohen (Anodot)

Average rating:

(4.40, 5 ratings)

Sequence to sequence (S2S) modeling using neural networks is increasingly becoming mainstream. In particular, it's been leveraged for applications such as speech recognition, language translation, and question answering. Arun Kejariwal and Ira Cohen walk you through how S2S modeling can be leveraged for the aforementioned use cases, visualization, real-time anomaly detection, and forecasting. Read more.

2:35pm–3:15pm Wednesday, September 11, 2019

Location: 230 C

Long-term real-time network traffic flow prediction using LSTM recurrent neural network

Secondary topics: Deep Learning, Machine Learning, Temporal data and time-series

Wei Cai (Cox Communications)

Real-time traffic volume prediction is vital in proactive network management, and many forecasting models have been proposed to address this. However, most are unable to fully use the information in traffic data to generate efficient and accurate traffic predictions for a longer term. Wei Cai explores predicting multistep, real-time traffic volume using many-to-one LSTM and many-to-many LSTM. Read more.

2:35pm–3:15pm Wednesday, September 11, 2019

Location: LL21 C/D

Fighting crime with graph learning

Secondary topics: Machine Learning

Mark Weber (MIT-IBM Watson AI Lab)

Average rating:

(4.00, 3 ratings)

Organized crime inflicts human suffering on a massive scale: upward of 700,000 people per year are "exported" in a $40 billion human-trafficking industry enslaving an estimated 40 million people. Such nefarious industries rely on sophisticated money-laundering schemes to operate. Mark Weber explores how a new field of AI called graph convolutional networks can help. Read more.

4:00pm–4:40pm Wednesday, September 11, 2019

Location: 230 C

Machine learning for autonomous driving: Recent advances and future challenges

Secondary topics: Computer Vision, Deep Learning, Machine Learning, Mobile Computing, IoT, Edge, Reinforcement Learning

Li Erran Li (Scale | Columbia University)

Tremendous progress has been made in applying machine learning to autonomous driving. Li Erran Li explores recent advances in applying machine learning to solving the perception, prediction, planning, and control problems of autonomous driving as well as some key research challenges. Read more.

4:00pm–4:40pm Wednesday, September 11, 2019

Location: LL21 C/D

Interpreting millions of patient stories with deep learned OCR and NLP

Secondary topics: Computer Vision, Deep Learning, Health and Medicine, Machine Learning, Mobile Computing, IoT, Edge, Text, Language, and Speech

Stacy Ashworth (SelectData), Alberto Andreotti (John Snow Labs)

Much business data still exists as challenging scanned or snapped documents. Stacy Ashworth and Alberto Andreotti explore a real-world case of reading, understanding, classifying, and acting on facts extracted from such image files using state-of-the-art, open source, deep learning-based optical character recognition (OCR), natural language processing (NLP), and forecasting libraries at scale. Read more.

4:50pm–5:30pm Wednesday, September 11, 2019

Location: 230 C

Named entity recognition at scale with deep learning

Secondary topics: Data, Data Networks, Data Quality, Deep Learning, Machine Learning, Text, Language, and Speech

Sijun He (Twitter), Ali Mollahosseini (Twitter)

Average rating:

(4.00, 1 rating)

Twitter is what’s happening in the world right now. To connect users with the best content, Twitter needs to build a deep understanding of its noisy and temporal text content. Sijun He and Ali Mollahosseini explore the named entity recognition (NER) system at Twitter and the challenges Twitter faces to build and scale a large-scale deep learning system to annotate 500 million tweets per day. Read more.

11:05am–11:45am Thursday, September 12, 2019

Location: Expo Hall 3

Using automated machine learning for hyperparameter optimization and algorithm selection

Secondary topics: Machine Learning, Temporal data and time-series

Francesca Lazzeri (Microsoft)

Average rating:

(3.50, 2 ratings)

Automated machine learning (AutoML) enables data scientists and domain experts to be productive and efficient. AutoML is seen as a fundamental shift in the way in which organizations can approach machine learning. Francesca Lazzeri outlines how to use AutoML to automate machine learning model selection and hyperparameter tuning. Read more.

11:55am–12:35pm Thursday, September 12, 2019

Location: Expo Hall 3

Semisupervised machine learning, the next frontier in AI

Secondary topics: Data, Data Networks, Data Quality

Vinay Rao (RocketML), Santi Adavani (RocketML)

Average rating:

(3.33, 6 ratings)

Current deep learning approaches require large amounts of labeled data. The creation of labeled data is expensive, error prone, and time consuming. Vinay Rao and Santi Adavani walk you through an effective learning method with minimum labelled data and human intervention. Read more.

1:45pm–2:25pm Thursday, September 12, 2019

Location: Expo Hall 3

PyTorch at scale for translation and NLP

Secondary topics: Text, Language, and Speech

Stef Nelson-Lindall (Facebook)

Average rating:

(2.00, 2 ratings)

PyText is a research to production platform that Facebook has leveraged to quickly develop state-of-the-art natural language processing (NLP) systems and deploy them to critical production use cases. Stef Nelson-Lindall explores several challenges with developing, training, and deploying real production systems with Torch, how to deal with them in NLP use cases, and more. Read more.

2:35pm–3:15pm Thursday, September 12, 2019

Location: 230 C

Transfer learning NLP: Machine reading comprehension for question answering

Secondary topics: Design, Interfaces, and UX, Machine Learning, Text, Language, and Speech

Anusua Trivedi (Microsoft)

Average rating:

(2.33, 3 ratings)

Modern machine learning models often significantly benefit from transfer learning. Anusua Trivedi details a study of existing text transfer learning literature. She explores popular machine reading comprehension (MRC) algorithms and evaluates and compares the performance of the transfer learning approach for creating a question answering (QA) system for a book corpus using pretrained MRC models. Read more.

2:35pm–3:15pm Thursday, September 12, 2019

Location: LL21 C/D

A practical guide toward explainability and bias evaluation in AI and machine learning

Secondary topics: Ethics, Security, and Privacy, Machine Learning

Alejandro Saucedo (The Institute for Ethical AI & Machine Learning)

Average rating:

(4.00, 3 ratings)

Alejandro Saucedo demystifies AI explainability through a hands-on case study, where the objective is to automate a loan-approval process by building and evaluating a deep learning model. He introduces motivations through the practical risks that arise with undesired bias and black box models and shows you how to tackle these challenges using tools from the latest research and domain knowledge. Read more.

4:00pm–4:40pm Thursday, September 12, 2019

Location: 230 A

Using deep learning models to extract the most value from 360-degree images

Secondary topics: Computer Vision, Deep Learning, Machine Learning

Shourabh Rawat (Zillow)

Average rating:

(4.00, 2 ratings)

Lately, 360-degree images have become ubiquitous in industries from real estate to travel. They enable an immersive experience that benefits consumers but creates a challenge for businesses to direct viewers to the most important parts of the scene. Shourabh Rawat walks you through how to identify and extract engaging static 2-D images using specific algorithms and deep learning methods. Read more.

4:50pm–5:30pm Thursday, September 12, 2019

Location: 230 C

Can behavioral analytics for enterprise security benefit from approaches in NLP?

Secondary topics: Ethics, Security, and Privacy, Machine Learning, Text, Language, and Speech

Ramsundar Janakiraman (Aruba)

While network protocols are the language of the conversations among devices in a network, these conversations are hardly ever labeled. Advances in capturing semantics present an opportunity for capturing access semantics to model user behavior. Ram Janakiraman explains how, with strong embeddings as a foundation, behavioral use cases can be mapped to NLP models of choice. Read more.