An ODSC West Guide to the Most Important Topics in Data Science Right Now
In this article, we’ll outline just a few of the most important topics in data science that our speakers will be presenting on at ODSC West Oct 29 - Nov 1 in San Francisco.
Every year, ODSC West gets hundreds of submissions from an array of incredibly talented data science practitioners. Those submissions offer a unique insight into what are some of the most important topics in data science right now. In this article, we’ll outline just a few of these topics that our speakers will be presenting on at ODSC West October 29th - November 1st.
Identified as one of the major threats to governments, politicians, businesses, and private individuals alike, deepfakes–both innocent and malevolent–have been dominating the news for several months. Originally only known as a side effect of GANs, deepfakes can create images, videos, and voice files capable of deceiving, at least initially, the general public. However, as the threat has escalated, so has the response from data scientists and tech organizations. Many are creating or encouraging the creation of tools that will help us identify and neutralize fake news and deepfakes. For more information on this topic, check out George Williams , Matt Zeiler, Mark Straub, Alex Comerford, Giorgio Patrini, Michael Tamir , and Jacob Baumbach’s ODSC West 2019 sessions.
As AI is increasingly used by industry and businesses to make decisions that affect peoples’ lives and future prospects (think job applications , loan applications , sentencing guidelines), it’s also increasingly important that we open the black box and understand how our AI models come to these decisions. Only by doing so will we be able to judge if a model is fair and if it leaves the possibility for those affected to make changes that can effectively influence the outcome. To learn more about this subject and develop the tools, skills, or knowledge that will help you improve the interpretability of your models, check out ODSC West 2019 sessions led by Mehrnoosh Sameki and Amy E. Hodler.
Although AI is largely used for fraud detection, AI models remain alarmingly susceptible to fraud themselves, particularly adversarial networks. Small changes to an image’s pixels can cause an AI model to mistake a picture of a Panda for that of a Gibbon. Or, of greater concern, an AI model can be tricked into thinking that an image of machine guns is, in fact, an image of a helicopter. As AI models are incorporated into our everyday lives, it becomes increasingly important to understand how to keep them secure. If this is a subject that interests you, don’t miss Dawn Song’s keynote address on AI and security at ODSC West this year.
As one of the major issues facing our planet today, climate change can feel like too big and too complicated of a problem to easily address. Fortunately, the data science community has shouldered this burden with gusto and is investigating possible applications of AI models, such as improving energy efficiency and helping us understand where and how climate change is currently having a direct effect. This year we’re explore how researchers are utilizing machine learning to address the effects of climate change with sessions led by speakers like David Rolnick .
Deep Reinforcement Learning:
In recent years deep reinforcement learning has been used to teach robots how to do household chores , beat competitive, world-class players at their games, and helped discover possible drug therapies of chronic diseases . At ODSC West, world-renowned Deep Reinforcement Learning expert Pieter Abbeel will be hosting a half-day tutorial on DRL during which he will cover the fundamentals of the topic and recent success stories.
Interrelated with some of the major topics we’ve already addressed, particularly explainable AI, bias continues to be a major concern in the data science and AI community. In addition to black box models making it difficult to discern which factors influence decision-making, limited training data lacking diversity is also a known contributor to bias in AI models. Fortunately, as the data science community has become aware of this problem, they have taken steps to address and remedy it. At ODSC West 2019, we’re excited to host speakers like Jakub Czakon , who will be discussing model fairness and bias.
Researchers and even ODSC presenters need to be mindful of this and strive to produce work that is accessible enough for others to understand, dissect, and build upon it. Fortunately, many organizations have recently stepped up to address this issue with tools and studies, such as NeurIPS’ Reproducibility Checklist and the Allen Institute for Artificial Intelligence’s Show Your Work study. At ODSC West, speakers such as Rajiv Shah and Paul Kowalczyk will be addressing reproducibility and its importance in their sessions. More importantly, our presenters know the importance of ensuring ideas, tools, and techniques outlined in talks and workshops are transferable once in your research and workplace.
In recent years, AI has begun to influence many aspects of the healthcare experience, including diagnosis, surgery, the development of new drugs, the use of big data, and treatment just to name a few. As AI is increasingly integrated into our healthcare experience, it’s important that we understand the many applications, as well as their challenges and opportunities. This year, ODSC West will have several speakers addressing these very issues, including Sriram Sankararaman, Andrew Long, and Kerry Weinberg.
The last, but definitely not the least, major topic that we’ll address in today’s article is homelessness. Although this is another complicated, intricate, and difficult to address societal issue, data scientists have been making progress in recent years. In 2018, researchers worked to build an algorithm that could help improve outcomes for those given housing assistance. And, in 2019 a company in Toronto built a chatbot to help the homeless community more efficiently access resources. At ODSC West, speakers such as Phebe Vayanos, PhD, will be addressing how to use AI to address homelessness.
Every year ODSC West strives to bring you content on the tools, platforms, and topics that influence and affect you every day. This year, we believe our talks, workshops, and trainings will cover content that is not only interesting and relevant, but also important for the creation of a better society and data science community. To learn more, and see how you might contribute to these efforts, we hope you’ll join us at ODSC West 2019 in just a few weeks.