Back to Events

Seminar Series

November 06

JP Morgan Tech Talk: Fair and explainable AI/ML for financial services

Jiahao Chen
JP Morgan

Part Of

12:00P
- 1:00P

Location

Star D463
Add to Calendar 2019-11-06 12:00:00 2019-11-06 13:00:00 America/New_York JP Morgan Tech Talk: Fair and explainable AI/ML for financial services Abstract: The financial services industry has many needs for fairness and explainability in artificial intelligence and machine learning, which stem from considerations of transparency, ethics, regulatory compliance, and risk management. [1] For example, banks must prove that the way that they approve mortgages comply with fair lending laws and promote community development, while at the same time managing risk appropriately. [2] These needs translate directly onto AI/ML solutions being developed for these business needs. In this talk, I introduce two research challenges that arise from the unique mix of business needs and regulatory constraints. First, we develop new methods for measuring bias in decision processes where labels for protected class membership cannot be observed [3, 4]. Second, we review the many definitions of fairness [5] and existing results on which definitions are mutually incompatible [6,7], and present our latest results exploring fairness-fairness and fairness-performance trade-offs.Wednesday November 6, 201912:00 PM- 1:00PMMIT CSAIL Stata Center, Star Conf. Room D463Lunch will be providedPlease register or email Callie Mathews- cmathews@csail.mit.edu to confirm attendance.https://www.eventbrite.com/e/csail-alliance-jp-morgan-tech-talk-tickets-79769983167

September 23

Dell EMC Tech Talk: Contextualized AI Models for Multiple Industries, and Building Relationships: Interactions and Graph Networks

Michael Shepherd and Ben Fauber
Dell EMC

Part Of

4:00P
- 5:00P

Location

Star D463
Add to Calendar 2019-09-23 16:00:00 2019-09-23 17:00:00 America/New_York Dell EMC Tech Talk: Contextualized AI Models for Multiple Industries, and Building Relationships: Interactions and Graph Networks As increasingly broader data domains are contextualized to enhance AI models, we find many industries are challenged to handle the sheer scale of this fast-moving technology shift. At Dell Technologies, our purpose is to drive human progress, through greater access to better technology, for people with big ideas around the world. Accomplishing this means we must be practitioners, researchers and leaders in our own right as AI takes a front row seat in everything we do. From manufacturing to predictive maintenance, supply chain optimization to fraud and sales to support we will provide an overview of how Dell Technology Services (DT Services) AI Research team is applying AI to industry wide problems. Michael will first discuss how CTO classifies projects as AI In, AI On, and AI For Dell and how Services manages and optimizes 65 Million devices with predictive models. Ben will then discuss the contextualization of data, through graph networks, and how graph modeling can enable novel applications, and present new challenges in data mining and scalable machine learning.SPEAKER BIOS: With 25 yrs experience, Michael Shepherd is a Distinguished Engineer and recognized technical evangelist who speaks globally on the impact of emerging technologies and AI. He currently serves as the Lead Technologist for the DT Services AI Research team. Michael sits on the Advisory Council for the University of Texas McCombs MSBA program and is on the core team for Xavier Health’s AI Initiative bringing FDA and industry together to advance AI in Healthcare. Michael has been granted thirteen hardware and software patents in eight countries. Ben Fauber is a senior data scientist at Dell Technologies in Austin, Texas, where he leads university research partnerships for the Applied Data Science Research team in DT Services. He earned his B.Sc. from Colorado State University and Ph.D. from The University of Texas at Austin. Prior to joining Dell Technologies, he held various scientific and leadership positions at Stanley Black & Decker, Genentech/Roche, AstraZeneca, and Pfizer. He is a co-inventor and author on more than 70 patents and peer-reviewed publications, including publications in the journals Nature, and Proceedings of the National Academy of Sciences. He currently serves on advisory boards for The University of Texas at Austin Dell Medical School and McCombs School of Business.Refreshments will be provided. Please RSVP or email cmathews@csail.mit.edu to confirm attendance. Star D463

April 23

Salesforce Tech Talk: Towards Versatile AI: Multi-task Learning and Generalization to New Tasks

Stephan Zheng
Salesforce Research

Part Of

12:00P
- 1:00P

Location

Star D463
Add to Calendar 2019-04-23 12:00:00 2019-04-23 13:00:00 America/New_York Salesforce Tech Talk: Towards Versatile AI: Multi-task Learning and Generalization to New Tasks Please register or email Callie Mathews cmathews@csail.mit.edu to confirm attendancehttps://www.eventbrite.com/e/salesforce-tech-talk-towards-versatile-ai-multi-task-learning-and-generalization-to-new-tasks-tickets-59182314897April 23rd, 12PM-1PM StarD463Lunch will be providedSalesforce Tech Talk: Towards Versatile AI: Multi-task Learning and Generalization to New TasksDeep neural networks have been top performers for machine learning problems on a single task, such as machine translation or playing Atari video-games. However, it is hard to achieve strong performance on multiple tasks simultaneously or generalization to unseen tasks, as data statistics and performance metrics can vary dramatically across tasks. In this talk, I will present two recent works that address these challenges.First, I will show decaNLP, a framework and competition for multi-task learning that unifies ten natural language processing (NLP) tasks as question-answering problems. This approach enables training unified models that can achieve competitive performance simultaneously on the ten NLP tasks, which include translation, summarization and text classification.Second, I will present how to derive theoretical generalization guarantees in reparametrizable reinforcement learning, in which trajectory distributions can be decomposed using the reparametrization trick. We theoretically derive and empirically verify Rademacher/PAC-Bayes generalization bounds for both intrinsic (due to overfitting within a single task) and external errors (due to shifts in world dynamics between tasks).Finally, I will give a high-level overview of machine learning research at Salesforce Research, including research on AI for Social Good and explainable AI.Presenter: Stephan Zheng, Research Scientist Salesforce ResearchBio:Stephan Zheng is a Research Scientist at Salesforce Research. He obtained his PhD in 2018 in the Machine Learning group at Caltech, advised by Professor Yisong Yue. His current research focuses on deep reinforcement learning in multi-agent environments. He has also worked on improving the robustness of deep learning and multi-resolution learning for spatiotemporal data.Previously, he received an MSc (Theoretical Physics) and BSc (Physics, Mathematics) from Utrecht University, read Part III Mathematics at the University of Cambridge and was a visiting student at Harvard University. He received the 2011 Lorenz Prize in Theoretical Physics from the Dutch Academy of Arts and Sciences, and was twice a research intern with Google Research and Google Brain.

April 17

STMicroelectronics Tech Talk: STMicroelectronics enables AI on the edge

Arcangelo Bruna
STMicroelectronics

Part Of

12:00P
- 1:00P

Location

Hewlett G882
Add to Calendar 2019-04-17 12:00:00 2019-04-17 13:00:00 America/New_York STMicroelectronics Tech Talk: STMicroelectronics enables AI on the edge Hewlett G882Wednesday April 17th, 12PM-1PM: Lunch will be providedPlease register or email Callie Mathews cmathews@csail.mit.edu to confirm attendance.https://www.eventbrite.com/e/tech-talk-stmicroelectronics-enables-ai-on-the-edge-tickets-59765404935STMicroelectronics enables AI on the edgeLeveraging the industry-leading position of its STM32 family of 32-bit Flash microcontrollers based on the Arm® Cortex®-M processor, STMicroelectronics is adding advanced Artificial Intelligence (AI) features to its design tools that enable fast and efficient embedded system implementation of trained neural networks. With this technical talk, STMicroelectronics will present the Company Vision for AI on the edge, give technical insights on the state of art and present the design tool with a /live demo/ on STM32 development boards.AI uses trained artificial neural networks to classify data signals from motion and vibration sensors, environmental sensors, microphones and image sensors, more quickly and efficiently than conventional handcrafted signal processing. ST's new neural-network developer toolbox is bringing AI to microcontroller-powered intelligent devices at the edge, on the nodes, and to deeply embedded devices across IoT, smart building, industrial, and medical applications.With STM32Cube.AI, developers can now convert pre-trained neural networks into C-code that calls functions in optimized libraries that can run on STM32 MCUs.Speaker: Arcangelo BrunaArcangelo Bruna joined STMicroelectronics in 1999, he holds a Master degree in EE and a PhD in applied mathematics. He is expert in Image Processing, Image coding/decoding and Image Analysis with Machine Learning, his developments have been applied in various fields such as automotive, robotics and SLAM for AR/VR. He is working on STMicroelectronics’ new developments on Machine Learning and CDNNs on the edge, and has enabled numerous advancement on AI for embedded applications. Arcangelo is managing several R&D projects with Universities and Research Centers in Europe and has held several lectures across various academic institutions. Arcangelo is senior staff engineer at ST, IEEE Senior Member and author of more than 80 papers and 31 patents.About STMicroelectronicsST is a global semiconductor leader delivering intelligent and energy-efficient products and solutions that power the electronics at the heart of everyday life. ST's products are found everywhere today, and together with our customers, we are enabling smarter driving and smarter factories, cities and homes, along with the next generation of mobile and Internet of Things devices.By getting more from technology to get more from life, ST stands for life.augmented.In 2018, the Company's net revenues were $9.66 Billion, serving more than 100,000 customers worldwide.

April 11

March 07

JP Morgan Tech Talk: Privacy Preserving Predictions

Antigoni Polychroniadou
JP Morgan

Part Of

12:00P
- 1:00P

Location

Star D463
Add to Calendar 2019-03-07 12:00:00 2019-03-07 13:00:00 America/New_York JP Morgan Tech Talk: Privacy Preserving Predictions While big data technology offers great promise, it introduces a challenge for privacy. Can we achieve the best of both worlds, extracting the benefits that big data offers while providing privacy to the data owner? Essentially, we would like to perform complex computations while preserving privacy. In this talk I am going to talk about privacy preserving computation and show how cryptography, streaming data and machine learning can support new ways of organizing predictive analytics. Presenter Bio:Antigoni Polychroniadou is a researcher at J.P. Morgan AI research and cryptographer lead in ROAR Data at J.P. Morgan. Antigoni was a junior Simons fellow, awarded by the Simons Society of Fellows, at Cornell Tech and a postdoctoral researcher in the Computer Science Department at Cornell University, hosted by Rafael Pass and Elaine Shi. Antigoni completed her Ph.D. at Aarhus University under the supervision of Ivan Damgård. She interned at IDC Herzliya, the Technion, University of California, Berkeley and IBM Research Thomas J. Watson. She holds an M.Sc. in mathematics of cryptography and communications from Royal Holloway University of London and B.Sc. in Computer Science and Economics from University of Macedonia, Greece.Please register OR email Callie Mathews, cmathews@csail.mit.edu to confirm attendance.

February 07

Nokia Bell Labs Tech Talk: Structure, Meaning, Action and Interaction - a future vision for augmenting human, machine, and network intelligence

Christopher White
Nokia Bell Labs

Part Of

12:00P
- 1:00P

Location

Star D463
Add to Calendar 2019-02-07 12:00:00 2019-02-07 13:00:00 America/New_York Nokia Bell Labs Tech Talk: Structure, Meaning, Action and Interaction - a future vision for augmenting human, machine, and network intelligence We sit at the cusp of a new revolution. A revolution born at the nexus of advances in devices used to sense ourselves and our environment, in algorithms that enable optimization, prediction, and learning and in networks that enable near instantaneous action and communication over terrestrial distances. This revolution promises to save time and create new knowledge through the automation of the mundane and augmentation of human, machine and network intelligence. Within this talk I will focus on several Bell Labs disruptive innovations that make possible this future and the technical challenges we use to define our research vision.Presenter: Chris White, Head Algorithms, Analytics and Augmented Intelligence Lab, Nokia Bell LabsChristopher A. White leads the Algorithms, Analytics & Augmented Intelligence (AAAI) lab in Nokia Bell Labs. He joined Bell Labs in 1997 after graduating with a Ph.D. in theoretical quantum chemistry from the University of California in Berkeley, California. His research interests include the development of computational models and methods for the simulation and control of interesting physical and digital systems. This has included work in areas ranging from linear scaling quantum chemistry simulations, to the design of new optical devices, to the global control of transparent optical mesh networks and to understanding and facilitating the propagation of ideas in organizations. In addition to the management of an international team of world-class researchers, Dr. White’s current work focuses on the creation of assisted thinking tools that leverage structural similarity in data with the goal of augmenting human intelligence. Lunch and light beverages provided

January 29

Apple IAP Tech Talk @ CSAIL Tuesday, January 29 from 12-1pm: Featuring Siri & HomePod Teams

Apple Siri and HomePod Engineers
CSAIL Alliances / Apple

Part Of

12:00P
- 1:00P

Location

Kiva/Patil Seminar Room (32-G449)
Kiva/Patil Seminar Room (32-G449)
Add to Calendar 2019-01-29 12:00:00 2019-01-29 13:00:00 America/New_York Apple IAP Tech Talk @ CSAIL Tuesday, January 29 from 12-1pm: Featuring Siri & HomePod Teams In two weeks from today on January 29th , engineers from Apple’s Siri and HomePod teams will be at CSAIL in the Kiva/Patil Seminar Room (32-G449) from 12-1pm to host a tech talk describing some of the new technologies Apple developed for their HomePod and Siri products.There will also be Apple recruiters available during and after the talk to speak with prospective students interested in working in the Siri teams at the company. If you would like to RSVP to speak with a recruiter about Apple opportunities while they are at CSAIL, you can do so on their MIT-Apple portal here (hyperlink). Otherwise, if you are only interested in attending the talk you are free to just show up and sign-in outside Kiva prior to the event as space will be limited. Food and refreshments from Flour will be also served. LINK TO RSVP: https://applecorp.avature.net/registration?projectId=11328&source=UR+Site+Portal&tags=mit%7Cspring_2019%7Cmassachusetts_institute_of_technology_tech_talk_with_siri_spring_2019%7Cprivacy+consent_obtained If you have direct questions for the Siri teams who will be at CSAIL that day, you may contact them at mit@apple.com . You may also feel free to contact me directly if you have any other questions pertaining to this event or CSAIL Alliances in general. Happy IAP! Kiva/Patil Seminar Room (32-G449)

December 10

BioMind Tech Talk: Improving the Accuracy and Quality of Healthcare with Machine Learning Applications

Evelyn Chee
BioMind

Part Of

12:00P
- 1:00P

Location

Kiva Room, 32-G449
Add to Calendar 2018-12-10 12:00:00 2018-12-10 13:00:00 America/New_York BioMind Tech Talk: Improving the Accuracy and Quality of Healthcare with Machine Learning Applications In this talk, we will discuss the different applications of machine learning in the medical field. In particular, how deep learning techniques have made the automation of complex medical tasks possible, ranging from medical image diagnosis to radiological report generation to radiotherapy treatment planning. Instead of detecting objects in natural images, could we implement R-CNNs to detect and classify abnormalities in medical images? Reinforcement learning is gaining popularity in the gaming field; would it be possible to apply the same idea to treatment planning systems to provide accurate treatment recommendations to doctors? Is GAN able to generate scans with isotropic voxels from low-resolution medical images? We will discuss how some of these techniques can impact the medical world in reducing the discrepancies in the quality of healthcare in rural and urban areas.Presenter: Ms. Evelyn Chee, Research Engineer, BioMindBio: Evelyn Chee graduated with a First Class Honours in Applied Mathematics from the National University of Singapore. She started her early work in deep learning back in 2016 on reinforcement learning. In 2018, she joined BioMind, which specialises in developing solutions using deep learning for the medical field. As a research engineer and team lead of the Radiotherapy Research Team, Evelyn has been deeply involved in fine-tuning BioMind’s medical image segmentation and registration processes.About BioMind: BioMind's mission is to improve efficiency and patient outcomes by empowering doctors and healthcare professionals with advanced technologies built upon deep learning techniques. The company built one of the world's first AI machines that can accurately diagnose brain tumors, vascular diseases, and stroke-related conditions. Lunch will be served. Please register to ensure we have an accurate catering order.

November 27

SenseTime Tech Talk: AI Innovation to Power the Future

Xiaogang Wang
SenseTime

Part Of

5:00P
- 7:00P

Location

Star Room, 32-D463
Add to Calendar 2018-11-27 17:00:00 2018-11-27 19:00:00 America/New_York SenseTime Tech Talk: AI Innovation to Power the Future Deep learning has penetrated the fundamental verticals that are intertwined with our daily lives. In the field of computer vision, SenseTime has been able to apply its globally-renowned DNN to smart cities, chips, autonomous driving, mobile phones, apps, and fintech. In this talk, SenseTime will discuss the foreseeable potential in these areas, the fundamental research necessary to further penetrate and disrupt, and the obstacles and questions that may reveal themselves along the way. Computer vision nowadays is not solely relegated to the realm of facial recognition but also involves the much more difficult task of a machine's ability to detect and recognize objects, movements, and environments. Acquisition of large sets of data, accuracy, security, and adoption still remain at the heart of making smart vision the next step in an industrial revolution. This presentation will provide some insight into SenseTime's research and commercialization capabilities and reveal why it is now the highest-valued AI unicorn in the world.The talk will run from 5-6pm and will continue with a networking session concluding at 7pm.About the speaker:Xiaogang Wang is the co-founder and Head of Research at SenseTime.About SenseTime:As the world’s highest-valued AI start-up company, SenseTime is well known for its proprietary algorithm in AI and its commercialization capabilities. SenseTime has formed the MIT-SenseTime Alliance as part of the MIT Quest for Intelligence, funding projects in all five schools at MIT. SenseTime has also collaborated and served over 700 well-known companies, including Honda, Qualcomm, China Mobile, UnionPay, Huawei, Xiaomi, OPPO, Vivo, Suning, etc.