Sign up or log in to bookmark your favorites and sync them to your phone or calendar.

Deep Learning [clear filter]
Saturday, October 20


Reducing AI biases by collecting balanced training data
The problem of bias in machine learning has raised significant concerns and become more critical than ever as AI technology rapidly spreads to all application domains.  While AI biases could occur due to the limitation of technology and design, majority of incidents happened because of the biased training data.  As AI powered systems are only as good as the data we feed them, skewed training data often result in AI models that show discrimination and amplify human prejudices.

Over 10 years enabling AI projects in the real world, Figure Eight’s platform has collected more than 10 billion human judgements, and has more than 100 million images labeled.  In an effort to prevent discrimination in AI, Figure Eight developed a framework that reduces biases in collecting training data in order to address the problem from the source.  By routing proper labeling tasks to proper labelers, our framework prevents potentially unwanted biases introduced by the labeling crowd.  The solution essentially performs a multi-objective optimization in order to achieve the desired distribution of labelers working on a given task.  In the talk, Monchu will share some use cases in Figure Eight where we applied the bias minimizer framework.  Empirical results will be shown to illustrate how the optimization process was done in the data collection process to minimize skewed training data.  The extension and implication to different scenarios and issues will then be discussed.

avatar for Monchu Chen

Monchu Chen

Director of Human-Computer Interaction, Figure Eight Technologies Inc.
Monchu Chen is the Director of Human-Computer Interaction at Figure Eight, the leading Human-in-the-Loop Machine Learning platform that transforms unstructured data into customized high quality training data.  He has over 15 years of experiences in practicing and researching human... Read More →

Saturday October 20, 2018 10:00am - 10:30am
Ballroom # 404


Leaders, Women, and Men Who Want Balance in Leadership
The financial cost of unhappiness at work is in excess of $550 billion per year because 68% of the American workforce is unengaged. Add to that increasing workplacehostility, discrimination, harassment, workforce division, miscommunication, and as of lately, the MeToo movement. All are evidence that changes in the way we invest in, train and engage our workforce are necessary. This old code is becoming too expensive to maintain.

But, what are we missing in our businesses? Deep employee dissatisfaction with our workplace cultures is pushing a new code to emerge.

Millennials, minorities, women, and men are questioning the current system’s effectiveness because it operates on an outdated code of workplace politics and hierarchies and it’s perceived sole focus on profits. This old code has not been updated to reflect the changing values that we expect of today’s business culture, which must include all people. Because of that, we are now at a breaking point.
There is a way to fix this. We can solve these problems by introducing a new code. The Women’s Code shows how to rebalance office culture. Innovation, creativity, and increased productivity and engagement happen when the focus shifts to building up people and their skills, developing strength-based leadership, and mapping out gender-neutral career paths. A workforce of individuals who feel recognized for their unique talents will produce up to 66% return on investment. Specifically, this new code will
- Show you what the old code is costing your business
- Skyrocket your people ROI
- Give you the five elements needed for diverse and equal team building
- Demonstrate how to master the essential three female-centric leadership attributes and how best you can use them in synergy with male-centric ones

Your takeaway is identifying how to take the first step that you can take immediately that will transform employee engagement and improve your overall financial performance.

avatar for Beate Chelette

Beate Chelette

Founder, The Women’s Code
Beate Chelette is The Growth Architect and a results-oriented businesswoman with an entrepreneurial spirit and a proven track record in growing, building, and scaling businesses. Once $135,000 in debt and a single mother, she successfully sold her business to a global entertainment... Read More →

Saturday October 20, 2018 10:30am - 11:00am
Ballroom # 404


AI and Deep Learning Trends in Genomics, Healthcare Informatics and Precision Medicine
The presentation will cover current trends in AI and Machine learning literature on Genomics , Healthcare informatics and our march towards precision medicine. This will also cover the ethics and other issues involved as we gather more and more of healthcare information about patients. There will be a few discussions on how AI can help doctors in patient management and the role of AI based prediction in patient care.

While the healthcare sector is being transformed by the ability to record massive amounts of information about individual patients, the enormous volume of data being collected is impossible for doctors to analyse. Machine learning comes to the rescue and provides a way to automatically find patterns and reason about data, which enables healthcare professionals to move to personalized care known as precision medicine.

As an illustration, healthcare providers can take advantage of machine learning to predict hospital re-admission for chronically ill patients. Greater support can be provided to patients identified most at risk of being re-admitted at discharge. By lowering the rate of readmission, we improve the lives of those most at risk and lower healthcare expenses. PatientMD team is involved in developing such algorithms.

Ninety three percent of healthcare information is in images. We will cover the latest in image recognition algorithms that our researchers at PatientMD are working on.

Finally, a introduction to chromputer model which uses CNNs sto predict histone modifications and chromatin states will be provided to audience interested in genomics.

Short Bio: Anirban Majumdar is the co-founder and CTO/COO of PatientMD. Prior to founding PatientMD, Anirban spent a decade in investments and finance. Anirban spent many years working in Silicon valley leading teams in companies like Oracle and BEA Systems.

Anirban is currently involved in rolling out PatientMD platform live in multiple countries.

Anirban is passionate about block chain, healthcare, genomics, mathematics, economics, finance, AI and computer modeling.

Anirban has BTech in Chemical Engineering from IIT Kharagpur, MS in Computer Science from Alabama and MBA in Analytical Finance from University of Chicago, Booth School of Business.

avatar for Anirban Majumdar

Anirban Majumdar

Co-Founder, CTO/COO, PatientMD Inc.
Anirban Majumdar is the co-founder and CTO/COO of PatientMD. Prior to founding PatientMD, Anirban spent a decade in investments and finance. Anirban spent many years working in Silicon valley leading teams in companies like Oracle and BEA Systems. Anirban is currently involved in... Read More →

Saturday October 20, 2018 11:00am - 11:30am
Ballroom # 404


Image Detection/ Object Recognition – Convolutional Neural Network
Our brains make vision seem easy. It doesn't take much effort for humans to tell apart her cousin and nephew, detect a face or recognize who it is. But these are actually hard problems to solve with a computer: they only seem easy because our brains are incredibly good at understanding images. Today, we will discuss the history of models used for such problems, what we've achieved so far and demonstrate combinations of state-of-art machine learning algorithms to tackle these problems and beyond.

avatar for Phillip Kim

Phillip Kim

Founder /Data Scientist, Banking & Quantitative Solutions
Phillip is a founder and data scientist at Banking & Quantitative Solutions LLC. He worked in various traditional & data analytics roles in legal, insurance, social media, finance, and investment banking. His expertise is in gathering and analyzing big data using various statistical... Read More →

Saturday October 20, 2018 11:30am - 12:00pm
Ballroom # 404


Making Blockchain Smart Contracts smarter with AI
One of the many benefits of Blockchain for Business is the automation of complex, high-value transactions within a business network. Smart Contracts is a feature of the Blockchain Platform that evaluates the mutually accepted terms and conditions between parties. Automating these transactions requires validating identity and proof of work to provide trust in the process. We will explore how Machine Learning, AI and Analytics can help make these Smart Contracts smarter.

avatar for Chris Tyler

Chris Tyler

Sr. Cloud Architect and Cognitive Evangelist, IBM
Chris Tyler is an IBM Technical Evangelist with over 25 years of experience in data, analytics and software development. He helps IBM Partners leverage IBM technologies to bring new products and services to market. Chris is an expert in embedded analytics and speaks to audiences around... Read More →

Saturday October 20, 2018 11:30am - 12:00pm
Ballroom # 403A + Conference Registration + Open talk


Panel Discussion #2 How AI Can Grow Business?
avatar for Mic Farris

Mic Farris

Director, KPMG
Mic serves as Director in KPMG’s Deal Advisory & Strategy Analytics practice and previously served as Vice President of Decision Science for CenturyLink. Mic is an experienced senior technology executive with over 20 years of proven success in creating, developing, and delivering... Read More →
avatar for Yutao He

Yutao He

Sr. Research Technologist, NASA/JPL
Dr. Yutao He is a Sr. Research Technologist and explorer at NASA Jet Propulsion Laboratory (JPL) leading technology development in innovative avionics systems for future space exploration. He is also an adjunct faculty member at UCLA and CSULA. Dr. He received his B.E. in Electrical... Read More →
avatar for Vladyslav Ivanov

Vladyslav Ivanov

Data Scientist
Vladyslav Ivanov is a data scientist with a proven record of developing and deploying data-driven solutions in the fields of Finance, IoT, and Education. His expertise includes Artificial Intelligence, Machine Learning, and Data Mining.
avatar for Marie Smith

Marie Smith

Marie Smith’s ambitious entrepreneurial spirit has led her to be a founding contributor in her 20 year career to nearly 50 media, tech and wellness companies and projects which have contributed to over $45 million in transactions. After bringing her talents to work at Warner Brothers... Read More →
avatar for Xiaolu Zhao

Xiaolu Zhao

CTO, Data Application Lab
Specialize in building the solutions for on-prem cloud computing and infrastructure, enterprise networking system and petabyte data warehouse. Automate and enhance IT operations by using analytics and machine learning to process data collected from various sources, spot and react... Read More →

Saturday October 20, 2018 1:00pm - 1:40pm
Ballroom # 404


Investment on AI
avatar for Shawn Chen

Shawn Chen

Global Partner, Plug and Play
Dr. Chen is the Global Partner of Plug and Play and the founder and chairman of Sias International University. He is also the Chairman of the Advisory Council of the Association of Universities of Asia and the Pacific (AUAP). He served as the President of the AUAP from 2014 to 2016... Read More →

Saturday October 20, 2018 1:40pm - 2:10pm
Ballroom # 404