Artificial Intelligence and Deep Learning are the current buzzwords of the tech industry. So, have you been wondering what all the fuss is about?
Have you wanted to incorporate AI into your products, but have no idea where to start? Are you letting terms like 'Big Data' or 'Differential Calculus Equations' intimidate you from being a part of the leading edge of this technology?
JOIN US for an all weekend, hands on workshop series Demystifying Deep Learning & AI. You will have the opportunity to gain insight & hear lighting talks from World Class Engineers in the field.
Learn how to use the latest AI open source platforms including:
SciKit | Tensor Flow | ConvNetJS | Keras
Network & Mingle with PhD students, faculty members, and industry research scientists. Share your projects and vision for the future of Deep Learning & Artificial Intelligence.
SATURDAY NOVEMBER 19TH, 2016 - JAMES IRVINE CONFERENCE CENTER
| Time | Description |
|---|---|
| 8:45 - 9:30 am | Check-in & Breakfast |
| 9:30 - 10:00 am | Opening Remarks |
| 10:00 - 11:00 pm | Intro & Advanced Sessions |
| 12:00 - 1:30 pm | Lunch is Provided |
| 1:00 - 5:00 pm | Intro & Advanced Sessions |
| 6:30 - 9:00 pm | Networking Reception @ Make Westing |
SUNDAY NOVEMBER 20TH, 2016 - JAMES IRVINE CONFERENCE CENTER
| Time | Description |
|---|---|
| 8:45 - 9:30 am | Check-in & Breakfast |
| 9:30 - 10:00 am | Opening Remarks |
| 10:00 - 12:00 pm | Intro & Advanced Sessions |
| 12:00 - 1:30 pm | Lunch is Provided |
| 1:00 - 5:00 pm | Intro & Advanced Sessions |
| 5:00 - 5:30 pm | Closing Remarks |
Python | Numpy | Scikit Learn | Tensorflow
Instructional videos for step by step installation guides:
Introduction to Machine Learning with Luis Serrano, PhD.
Abstract
INTRO TO DEEP LEARNING
A friendly and pictorial dive into solving classification problems with deep learning. Topics will include logistic regression, gradient descent, and deep neural networks.
Demystifying Deep Learning Presentation
A Friendly Introduction to Deep Learning Slides
Numpy, Scikit Learn, & Intro to Deep Learning using Tensorflow with Abhishek Sharma
Abstracts
HANDS ON ML
Learn to use ML algorithms in your applications. In this session we will cover basics of numpy (a python library used for mathematical operations) and then move on to implementing ML algorithms using sci-kit learn.
HANDS ON TENSORFLOW
This session will cover the basic architecture of Tensorflow followed by implementation of basic deep learning architectures for classification tasks.
NOTE
Please make sure that numpy and scikit learn and Tensorflow are installed in your machine before coming for the sessions. Instructions for the same will be shared soon.
Machine Learning & Deep Learning Presentation
Intro to Deep Learning for Images in Keras with Stephan Egly & Malaikannan Sankarasubbu
DEEP LEARNING FOR IMAGES FROM HOLBERTON SCHOOL
Understand the basics of deep neural networks and play with Keras, a minimalist, highly modular neural networks library, written in Python and capable of running on top of either TensorFlow or Theano.
Deep Learning with the Holberton School
Intro into Keras and Image Classification
Internal workings of a convnet and the process of implementing it on Spark with Jeremy Nixon
Abstract
ELEGANCE OF CNNS IN PRACTICAL APPLICATIONS
Convolutional Neural Networks are as elegant as they are tremendously successful in practical applications. This talk will demonstrate how they operate at a low level, and then provide a number of diverse viewpoints on interpreting their performance. Those views will include manifold learning, representation learning, automating feature engineering and more.
Understanding Convolutional Neural Networks Slides
Overfitting and regularisation in Machine learning with Dmytro Lituiev
Abstract
OVER-FITTING IN DEEP LEARNING
Over-fitting, a phenomenon when a model explains well the data set it has seen, but fails to generalize to the new data is common to both simple linear models and complex deep neural networks. This workshop will provide intuition of such behavior and discuss ways to prevent it. In particular, we will focus on a weight penalization as a way to improve generalization. We will implement and train a model and look at it from probabilistic and linear algebra perspective.
Overfitting and Regularization
Classifying images using DCNN in Keras with Sujit Pal
Abstract
DCNNS IN KERAS
This talk will describe how a Deep Convolutional Network (DCNN) trained on the ImageNet dataset can be used to classify images in a completely different domain. The intuition that the training process teaches the DCNN to extract good features from images is explored with visualizations. Transfer Learning freezes the bottom layers of the DCNN to extract image vectors from a training set in a different domain, which can then be used to train a new classifier for this domain. Fine tuning involves training the pre-trained network further for the target domain. Both approaches are demonstrated using a VGG-16 network pre-trained on ImageNet to classify medical images into 5 categories. Code examples are provided using Keras.
Transfer Learning and Fine Tuning for Cross Domain Image Classification with Keras Presentation
Supplement Answer to how Backpropogation works across a max/min pool layer
Deep Learning for Recommendation Systems with Rumman Chowdry
Abstract
DEEP LEARNING FOR RECOMMENDATION SYSTEMS
This talk will cover an overview of different types of recommender systems, with a focus on using deep learning for recommendation systems. Notable recent application areas are music recommendation, news recommendation, and session-based recommendation.
Demystifying Recommendation Systems
In depth look at Word2Vec with Andy Zhang
Abstract
DEEP LEARNING MATH BY EXAMPLE
A case study with word2vec, a commonly used technique for pre-training word embeddings. What is the training process doing, what is being learned, and why do they work? The intuition gained will be useful for thinking about how language models with embeddings are trained in general.
Word2Vec Deconstructed (and Simplified) Slides
Word2Vec Workshop Lecture Notes
Exploding / Vanishing Gradient Problem with Alex Shim
Abstract
DEEP DIVE INTO THE VANISHING/EXPLODING GRADIENT PROBLEM
This talk is an in-depth explanation of Long Short-Term Memory and how they solve the vanishing/exploding gradient problem.
The exploding and vanishing gradient problem and LSTMs Slides
Did Big Data Fail us in the Presidential Election? with Rumman Chowdhury
Abstract
DID BIG DATA FAIL US IN THE PRESIDENTIAL ELECTION?
Not a single major poll, megapoll, or survey in this election successfully predicted the outcome. As a result, an already skeptical American population is experiencing a big data backlash. What happened? Is our faith in big data misguided? In this short talk, I discuss how the misconceptions in the data science community about the accuracy of models and polling, combined with a basic lack of understanding and explanation in the public about statistical models and their outcomes, led to arguably the biggest Big Data debacle since Google Flu Trends.
Using convolutional neural networks to classify Monet Paintings with Samuel Bozek
Abstract
CNN'S FOR FINE ART CLASSIFICATION
This talk will describe the use of convolutional neural networks to classify Monet Paintings.
Developing chatbots with AI with Masha Kubyshina
Abstract
CHATBOTS
Hear a summary of interviews with bot developers, platform developers, clients and users to understand motivation of each segment. This talk will cover the problems facing each sector of the bot ecosystem.
Learn about platforms that can be used by companies to build their bots and what marketing tools are available help advance bot discovery.
Incorporating ML into Robotics & Computer Vision with Carlos Uranga
Abstract
MACHINE LEARNING FOR ROBOTICS
What is a robot? Why is A.I. important in personal robotics? We now have the ability to digest enormous amounts of data, to read our own thoughts, and are now able to prevent disease - rather than treating it. The era of personalized robotics is upon us, designed for you, better serving us as individuals. This next era of robotic automation, brings them out of the factories ( big industry), and into our everyday lives; much of this, will be enhanced and accelerated by A.I.
Sam Bozek is a Canadian Chemist who loves alliteration and art. He decided to use computers to help with the art thing, after being a chemist ruined too many pant legs.
Rumman is a Senior Data Scientist at Metis and PhD Candidate in Political Science at the University of California, San Diego. Masters in Quantitative Methods from Columbia University and undergraduate degrees from MIT.
Her diverse background in both academia and the private sector emphasizes providing actionable and scaleable solutions using advanced statistical and machine learning techniques. She has successfully led industry-awarded data science teams, educates budding data scientists, and continues to help Fortune 500 companies leverage their data science capabilities.
She uses her study of human behavior in forecasting and the social sciences to inform her work in Data Science.
Stephane is passionate about solving problems that are important to society and social progress. He has a uniquely varied background, having worked on four different continents, as a Space Operations Engineer, Professor of Applied Physics, and Software Entrepreneur. As an AI Scientist at datalog.ai, his current obsession is modeling the acoustics of human speech for Deep Learning applications.
Masha's background is in Marketing, Customer Development and Linguistics. She started working on a personalized English learning tool with AI backend in 2014. In fall this year her team developed a bot to help people improve their English.
Dima Lituiev does human genetics research at the UCSF, co-organizes a Deep Learning Study Group, and occasionally teaches data science. Born in Ukraine, Dima did his PhD at the University of Zurich, Switzerland, where he analysed images and built model of development of plant flower parts. His interests are in probabilistic models, machine learning, genetics, and developmental biology.
JEREMY NIXON - SPARK TECHNOLOGY CENTER
Jeremy is a machine learning engineer dedicated to understanding the principles underlying information and intelligence and deploying them in meaningful real world applications. He has been working on deep learning in Apache Spark's MLlib since he graduated from Harvard in Applied Mathematics and Computer Science.
Sujit builds intelligent systems around research content. His primary interests are IR, NLP, ML, DL, ontologies and distributed processing. Prior to this, he built ontology backed semantic search, contextual advertising and EMR analytics for consumer healthcare.
MALAIKANNAN SANKARASUBBU - DATALOG.AI
Malai is the Founder and CTO at Datalog.ai, where he loves to take on the challenges of unstructured data. He actively works on Natural Language Processing and Computer Vision problems, with applications in conversational agents and health. He loves high performance computing, and trains Deep Learning models on GPUs to keep his home warm in the winter.
Luis Serrano is the Machine Learning Nanodegree Lead at Udacity. Previously he was a Machine Learning Engineer at Google for 2 years, working on the video recommendations at YouTube. Before that, he was an academic, and was working as a postdoctoral fellow in Mathematics at the University of Quebec at Montreal. Luis obtained his PhD in Mathematics at the University of Michigan, and his Bachelors and Masters in Mathematics at the University of Waterloo.
Organizer of Deep Learning Enthusiasts Meetup Group and a backend engineer at Metamind. He received his Master’s in Computer Science (AI track) from University of Illinois at Urbana Champaign (UIUC). His interests include Deep Learning, Natural Language Processing and large-scale Machine Learning.
He worked with Prof. ChengXiang Zhai, UIUC, on discovering controversial points in news articles by leveraging social media conversations. He was the recipient of the Cognitive Science/Artificial Intelligence Award – 2014, awarded by the Computer Science Department, UIUC.
ALEX SHIM - RESEARCH CONSULTANT, DATA SCIENTIST
Alex studied theoretical mathematics at the California Institute of Technology, then went into educational research. He has worked as the last few years as a research consultant for academic resesrch and data scientist consultant for businesses, and is now specializing in deep learning research.
CARLOS URANGA - SINGULARITY UNIVERSITY
Carlos' most recent endeavors are using Deep Learning (A.I.) in two of his startups (Biofabs, LUSH Robotics) - coupling algorithms and massively parallel computing to mimic the very neural networks that govern the behavior of what our brains do best - pattern recognition. Carlos is a serial entrepreneur, leading and advising startups as founder/ developer/ product manager/ instructor. His passion is one of leveraging converging disciplines, and harnessing the power of distributed innovation & digital fabrication - and putting them to work in robotics. Most recently he was the Director of the Innovation Lab ( iLab) at Singularity University (SU) where he served the mission to educate and empower present and future leaders towards leveraging exponential technologies.
Carlos has always prescribed to the SU mindset, in that everyone should inspire to ‘think big’ as an entrepreneur, especially when subscribing to make things that will help a billion people within 10 years. Carlos holds a Masters in Bioengineering from UCSD and the equivalent of a Masters in Computer Vision, and Robotics from INPG/ INRIA in Grenoble, France.
Self-taught software engineer with wide variety of experiences in frontend, mobile, distributed systems, data engineering, and machine learning. Avid student of new deep learning techniques and excited about applying recently published methods to applications in NLP and gaming.