Deep Dive into TensorFlow

Due to security concerns we need the First and Last name, email and company name. Please register using EVENTBRITE.COM (https://www.eventbrite.com/e/san-francisco-meetup-deep-dive-into-tensorflow-tickets-26537575603)

Many thanks to Autodesk (http://www.autodesk.com/) for hosting the Tensorflow meetup!

Agenda:

6:30 - Doors open. Networking. Members meet each other. Pizza

7:00 - Visual Search Engine using TensorFlow by Delip Rao, Founder @ Joostware

7:30 - Q&A break.

7:45- Using Deep Q Networks to Learn Game Strategies by Akshay Srivatsan, Software Engineering Intern

8:30 - Q&A break.

8:45 - Wrap-up.

DETAILED AGENDA:

Visual Search Engine using TensorFlow

Speaker: Delip Rao, Joostware

Delip Rao is the founder of Joostware, a San Francisco based company, that specializes in consulting and building highly scalable, bespoke solutions involving machine learning, search, and natural language processing. Delip has worked on NLP and ML research problems involving semi-supervised learning, graph based ranking, sequence learning, distributed machine learning, and more, and has published several highly-cited papers in these areas. At Joostware, Delip works 1-1 with his clients (early-mid stage startups) to help them set up their big data infrastructure, identify and implement novel machine learning, search, and natural language processing based solutions, and training on machine learning. Prior to founding Joostware, Delip worked on building ML/NLP research & products at Amazon, Twitter, and Google Research.

Using Deep Q Networks to Learn Game Strategies

Reinforcement learning is a process by which an agent can learn to interact with its environment in a specific way based on positive and negative feedback it receives from its interactions. In this talk, we will discuss a deep learning model recently developed by Minh et al 2015 to learn optimal control patterns from visual input using reinforcement learning. This technique is highly generalizable and is capable of achieving better than human performance in several specific video game environments. After discussing the theory, we will demonstrate how to implement this model using the TensorFlow library. We will also discuss more recent advancements in reinforcement learning such as Google's AlphaGo, which famously defeated Lee Sedol in March of this year. While these tasks are narrowly scoped, the success these techniques have had demonstrates the potential for deep learning as a powerful and generalizable method for learning high-level control schemes.

Speaker: Akshay Srivatsan, Software Engineering Intern

Akshay has a strong interest in machine learning, specifically in natural language processing, and has experience with graphical model approaches such as LDA and also deep learning based techniques such as word vector embeddings. Recently Akshay has become interested in TensorFlow as an effective toolkit for deep learning. Using the library, he has implemented projects on deep Q learning, and on deep canonical correlation analysis.

Due to security concerns we need the First and Last name, email and company name. Please register using EVENTBRITE.COM (https://www.eventbrite.com/e/san-francisco-meetup-deep-dive-into-tensorflow-tickets-26537575603)