ACM Data Science Camp 2017, Silicon Valley
Event Information
Description
Overview:
Data Science Camp is SF Bay ACM's annual event combining sessions, keynote, and optional tutorial (extra-fee). It's an excellent opportunity to learn about Data Science and connect with others, and we keep it near-free ($10 charge, includes coffee & lunch), now running in its ninth year.
See http://www.sfbayacm.org/data-science-camp-2017/ for a detailed description of this event.
Additional session topics are invited and may include (but are not limited to):
- Models: Deep Learning, xgboost, clustering, training, deployment, feature engineering
- Domains/verticals: Big Data, e-commerce, fraud detection, search, NLP/ontologies, trading/finance, Bitcoin, IT security, healthcare, environmental
- Tools and technologies: Spark/MLib, GPU, R, Python, PMML, Hadoop
- Related areas: Visualization, Data Engineering, Career Opportunities, Hiring Roundtable et al.
Session proposals are welcome from both individuals and companies. Please consider volunteering to speak or recommending people to give talks. Submit or vote on sessions on the "Session Proposals" tab here.
Schedule:
- 8:00 am – 8:40 Arrive, register for class, network, coffee
- 8:40 am – 10:40 Class: AI and Deep Learning with Python and Keras ($60, includes full day)
- (Class details are below)
- 10:30 am – 11:00 People coming for just the Camp ($10) arrive, register and network (past talks below)
- 11:00 am Camp Kickoff
- Major Sponsor 5 min presentations
- Keynote Presentation, 50 min "Data Science: Let’s Cut the Hype & Measure its Value to Prevent a Y2K Like Fiasco" by Alo Ghosh, A professor → advisor → startup → PE → VC guy’s perspective
- 12:25 Session Proposals (30 sec description, count audience hands, assign to a room for that sized audience)
- 1:15 Lunch, post Session Matrix (4 time slot rows by 4-7 room columns)
- 2:00 – 2:50 Session 1 (over all the rooms used, likely subdivide the main room that seats 410)
- 3:00 – 3:50 Session 2
- afternoon coffee and snacks
- 4:00 – 4:50 Session 3
- 5:00 – 5:50 Session 4
- 6:00 – 6:30 Session Summary, in the largest part of the main room
- 6:45 All audience should be out of the building
Morning Class: AI and Deep Learning with Python and Keras
Time and Cost:
- The class will be from 8:40 – 10:40am, with a $60 charge which includes the afternoon camp. For just the Camp, from 10:40am and the rest of the day is $10 or less.
Speaker:
- Bhairav Mehta is Senior Data Scientist at Apple and founder of DataInquest, which gives data science training
What Will I Learn?
- To describe what Deep Learning is in a simple yet accurate way
- To explain how deep learning can be used to build predictive models
- To distinguish which practical applications can benefit from deep learning
- To install and use Python and Keras to build deep learning models
- To apply deep learning to solve supervised and unsupervised learning problems involving images, text, sound, time series and tabular data.
- To build, train and use fully connected, convolutional and recurrent neural networks
- To look at the internals of a deep learning model without intimidation and with the ability to tweak its parameters
- To train and run models in the cloud using a GPU
- To estimate training costs for large models
- To re-use pre-trained models to shortcut training time and cost (transfer learning)
Requirements
- Knowledge of Python, familiarity with control flow (if/else, for loops) and pythonic constructs (functions, classes, iterables, generators)
- Use of bash shell (or equivalent command prompt) and basic commands to copy and move files
- Basic knowledge of linear algebra (what is a vector, what is a matrix, how to calculate dot product)
- Use of ssh to connect to a cloud computer
Description
- This training is designed to provide a introduction to Deep Learning using Keras. It is aimed at beginners and intermediate programmers and data scientists who are familiar with Python and want to understand and apply Deep Learning techniques to a variety of problems.
- We start with a review of Deep Learning applications and a recap of Machine Learning tools and techniques. Then we introduce Artificial Neural Networks and explain how they are trained to solve Regression and Classification problems.
- We introduce and explain several architectures including Fully Connected, Convolutional and Recurrent Neural Networks, and for each of these we explain both the theory and give plenty of example applications.
- The goal is to provide students with a strong foundation, not just theory, not just scripting, but both. At the end of the course you’ll be able to recognize which problems can be solved with Deep Learning, you’ll be able to design and train a variety of Neural Network models and you’ll be able to use cloud computing to speed up training and improve your model’s performance.
- Software engineers who are curious about data science and about the Deep Learning buzz and want to get a better understanding of it
- Data scientists who are familiar with Machine Learning and want to develop a strong foundational knowledge of deep learning
Agenda:
- Machine Learning and Deep Learning Introduction
- Gradient Descent
- Neural Network
- Convolution Neural Networks
- Recurrent Neural Network
- Keras Intro along with other libraries like Theano, Tensorflow, DL4J, MXnet etc.
- Set up your environment and AWS powered GPU based Jupyter Notebook
- Identify Keras and other Deep Learning libraries are installed
- Load image data from MNIST and CIFAR.
- Preprocess input data for Keras.
- Preprocess class labels for Keras.
- Define model architecture.
- Compile model.
- Fit model on training data.
- Evaluate model on test data.
- Load Time Series data from Airline Passenger Data
- Define RNN model Architecture
- Fit RNN model
- Evaluate model on test data
- Improve Performance
- Track performance
- Some demonstrations of AI based technologies
Speaker Bio:
- Bhairav Mehta is Senior Data Scientist with extensive professional experience and academic background. Bhairav works for Apple Inc. as Sr. Data Scientist.
- Bhairav Mehta is experienced engineer, business professional and seasoned Statistician / programmer with 19 years of combined progressive experience working on data science in electronics consumer products industry (7 years at Apple Inc.), yield engineering in semiconductor manufacturing (6 years at Qualcomm and MIT Startup) and quality engineering in automotive industry (OEM, Tier2 Suppliers, Ford Motor Company) (3 years). Bhairav founded a start up DataInquest Inc. in 2014 that is specialized in training/consulting in Artificial Intelligence, Machine Learning, Blockchain and Data Science.
- Bhairav Mehta has MBA from Johnson School of Management at Cornell University, Masters in Computer science from Georgia Tech (Expected 2018), Masters in Statistics from Cornell University, Masters in Industrial Systems Engineering from Rochester Institute of Technology and BS Production Engineering from Mumbai University.
- https://www.linkedin.com/in/mehtabhairav/
- http://www.datainquest.com/
Keynote Session
"Data Science: Let's Cut the Hype & Measure its Value to Prevent a Y2k Like Fiasco" from Alo Ghosh, a professor --> advisor --> startup --> PE --> VC guy's Perspective
Talk Description:
From ‘data deluge’ to ‘deep learning’, data science has been hyped to the point that businesses across the spectrum have spent billions, spurred on mostly by ‘FOMO’, only to find very little to show for this spend. There is today widespread CXO frustration re data science. To prevent a Y2K like fiasco, the data science community must show value, not from trumped up analyst reports and pithy product demos, but by addressing their key pain points:
- Data prep/munging/wrangling solutions, particularly regarding unstructured data,
- REAL data science talent and its ubiquitous unavailability, despite its myriad credentials,
- Expose machine learning (pattern/anomaly detection) and its celebrity child – deep learning (“the only real success of deep learning so far has been the ability to map space X to space Y using a continuous geometric transform, given large amounts of human-annotated data”) for what their minuses and pluses really are. True learning must combine data with rules.
- Black boxes created by ML/DL algorithms deter their use in financial and medical worlds
And quickly moving on to demonstrate how it can add real, measurable economic value to businesses:
- Strategy – Harness established tools like ‘DCF, real option and game theories’ to draw valuations of strategic alternatives and then monitor performance of the chosen strategy
- Finance – Marshal internal and market data to track company valuations and risk premiums
- Marketing – Track brand valuation and use dynamic pricing to capture economic surplus
- Supply Chains – Emulate what Amazon does so well
Greed rather than fear will establish data science as businesses’ primary value adding spend.
Keynote Background Reading (optional):
“The Dark Secret at the Heart of AI” MIT Technology Review, April 11, 2017
“How CEOs Can Keep Their Analytics Programs from Being a Waste of Time” Harvard Business Review, July 21, 2016
“The Age of Analytics: Competing in a Data-Drivin World,” McKinsey & Company, 2016
Alo Ghosh’ Bio:
Based on my expertise (PhD Wharton + 4 Master’s) and 35-yr. experience (Wharton professor, McKinsey NYC finance expert, PE & VC partner in NY & ASEAN, $4B country fund head for 3 yrs, bootstrapped Silicon Valley fintech startup to $100M in the 1990’s), I will provide an overview of the real-life footprint of data science and a proven way to measure its impact first developed at McKinsey.
MBA-PhD (Finance-Wharton). https://www.linkedin.com/in/draloghosh
Hands-on Silicon Valley co-founder, co-funder, interim CXO and risk manager of FinTech startups in Insurance & Investments, such as AI Labs
- Co-founder, interim CXO, seed funder to Silicon Valley FinTech & EdTech startups.
- Advisor to decision makers in hands-on creation of sustainable shareholder value.
- Trained data science expert in statistics-econometrics, O.R., quantitative finance.
- Created business plans for projects and ventures securing $’000 million in funding.
- Sourced large PE deals, headed country wealth fund, consulted widely in S/E Asia.
- Taught Wharton MBA, co-led McKinsey finance, forged global fintech consultancy.
Thirty-five years of learning & practice at the cutting edges of finance, strategy, technology in some of the world’s most storied institutions as well as with the most diverse of startups (including my own), universities, governments, private equity funds, hedge funds, investment banks & sustainable development non-profits/NGOs in several different parts of the world.
Example Sessions From Prior Years
Session Proposal Process
- Proposing sessions in advance on the "Session Proposals" tab is encouraged, to attract people to attend specific talks. The voting on the web gives guidance on popularity of a talk and gives the audience an idea of their 4 favorite talks. There are 4 time slots, starting on the hour, from 2:00 to 5:00. On the web, there is both a “thumbs up” and “thumbs down”. We invite audience members to keep the number of thumbs up to 4-6 (including any talks you give).
- The day of the event, Saturday 10/14/2017, at 12:25, we will go through a live 30 second session proposal from the presenters in the room. Then we ask the audience for a show of hands for room sizing for the given proposal (i.e. for a room that seats 40 or 120). At the END OF LUNCH, the organizers set up the final session matrix and post it on this web site, at the top of the "Session Proposals" tab.
- For a number of years, we have hosted talks presented from another country. If you are remote or in another country, make arrangements to have your talk proposed by someone in person using Skype, Webex, Zoom or your favorite screen sharing. That person would go to your session with a computer with your remote presentation setup to facilitate the microphone to the audience and to help facilitate any audience questions to the speaker.
2016 Sessions
2015 Sessions
Word Cloud image was from here
