Events

Hide Filters Show Filters
Hide Filters Show Filters

Event Type

Impact Area

Research Area

Seminar Series

Reset Filters

July 06, 2022

No events scheduled

July 07, 2022

Jiasi Shen Thesis Defense: Program Inference and Regeneration via Active Learning

Jiasi Shen
MIT-CSAIL
1:00P
- 2:00P

Location

32-G575
Add to Calendar 2022-07-07 13:00:00 2022-07-07 14:00:00 America/New_York Jiasi Shen Thesis Defense: Program Inference and Regeneration via Active Learning Title:Program Inference and Regeneration via Active LearningAbstract:Software now plays a central role in numerous aspects of human society. Current software development practices involve significant developer effort in all phases of the software life cycle, including the development of new software, detection and elimination of defects and security vulnerabilities in existing software, maintenance of legacy software, and integration of existing software into more contexts, with the quality of the resulting software still leaving much to be desired. The goal of my research is to improve software quality and reduce costs by automating tasks that currently require substantial manual engineering effort.I present a novel approach for program inference and regeneration, which takes an existing program, learns its core functionality as a black box, builds a model that captures this functionality, and uses the model to generate a new program. The new program delivers the same core functionality but is potentially augmented or transformed to operate successfully in different environments. This research enables the rejuvenation and retargeting of existing software and provides a powerful way for developers to express program functionality that adapts flexibly to a variety of contexts. In this talk, I will show how we applied these techniques to database-backed programs and discuss the broader implications of these techniques.Bio:Jiasi Shen is a Ph.D. candidate at MIT EECS.Thesis Committee:Martin Rinard, Armando Solar-Lezama, Saman Amarasinghe 32-G575

July 08, 2022

THESIS DEFENSE: Machine Learning Approaches For Equitable Healthcare

Irene Chen, MIT-EECS
11:00A
- 12:00P

Location

Star (32-D463) and Zoom: https://mit.zoom.us/j/2735218633
Add to Calendar 2022-07-08 11:00:00 2022-07-08 12:00:00 America/New_York THESIS DEFENSE: Machine Learning Approaches For Equitable Healthcare Abstract: Advances in machine learning and the explosion of clinical data have demonstrated immense potential to fundamentally improve clinical care and deepen our understanding of human health. However, algorithms for medical interventions and scientific discovery in heterogeneous patient populations are particularly challenged by the complexities of healthcare data. Not only are clinical data noisy, missing, and irregularly sampled, but questions of equity and fairness also raise grave concerns and create additional computational challenges. In this thesis, I present novel approaches for leveraging machine learning towards equitable healthcare. The thesis concludes with a discussion about how to rethink the entire machine learning pipeline with an ethical lens to building algorithms that serve the entire patient population.Thesis Committee: David Sontag (MIT), Peter Szolovits (MIT), Marzyeh Ghassemi (MIT) Star (32-D463) and Zoom: https://mit.zoom.us/j/2735218633

July 13, 2022

July 14, 2022

Thesis Defense: Balancing Exploration and Exploitation: Loss-Targeted Exploration for Scientific Decision-Making

Genevieve Flaspohler
CSAIL MIT
10:00A
- 11:30A

Location

32-G882 (Hewlett Room) Zoom Hybrid (Register here: https://mit.zoom.us/meeting/register/tJEpd-6srT4qGdWKputtTO3eVxeeFEDUIr25)
Add to Calendar 2022-07-14 10:00:00 2022-07-14 11:30:00 America/New_York Thesis Defense: Balancing Exploration and Exploitation: Loss-Targeted Exploration for Scientific Decision-Making Abstract:How do we collect observational data that reveal fundamental properties of scientific phenomena? This is a key challenge in modern scientific decision-making. Scientific phenomena are complex---they have high-dimensional and continuous state, exhibit chaotic or complex dynamics, and generate incomplete or noisy sensor observations. This thesis argues that autonomous decision-making in real-world, scientific domains requires loss-targeted exploration---exploration strategies that are tuned to a specific task or loss function. By explicitly quantifying the change in task performance due to exploratory actions, we enable decision-makers that explore parsimoniously and contend with highly uncertain real-world environments. We develop novel algorithms for loss-targeted exploration in partially observable Markov decision processes (POMDPs) and online learning problems. These methods are motivated by and applied to real-world scientific applications, including robotic deep sea hydrothermal plume discovery and climate and weather forecasting. This thesis demonstrates that autonomous decision-making can enhance human scientific discovery, placing sensors in the right place at the right time to validate a hypothesis or collect a critical observation.Thesis Supervisor(s): John Fisher and Nicholas RoyCommittee: John Fisher (MIT), Nicholas Roy (MIT), Claudia Cenedese (Woods Hole Oceanographic Institution), Munzer Dahleh (MIT), and Lester Mackey (MSR New England) 32-G882 (Hewlett Room) Zoom Hybrid (Register here: https://mit.zoom.us/meeting/register/tJEpd-6srT4qGdWKputtTO3eVxeeFEDUIr25)

The Effect of Epigenetic Blocking on Multi-Objective Genetic Algorithms

Sizhe Yuen
University of Southampton
4:00P
- 5:00P

Location

32-G449, Patil/Kiva
Add to Calendar 2022-07-14 16:00:00 2022-07-14 17:00:00 America/New_York The Effect of Epigenetic Blocking on Multi-Objective Genetic Algorithms Hundreds of Evolutionary Computation approaches have been reported. From an evolutionary perspective they focus on two fundamental mechanisms: cultural inheritance in Swarm Intelligence and genetic inheritance in Evolutionary Algorithms. Contemporary evolutionary biology looks beyond genetic inheritance, proposing a so-called "Extended Evolutionary Synthesis". Many concepts from the Extended Evolutionary Synthesis have been left out of Evolutionary Computation as interest has moved towards specific implementations of the same general mechanisms.One such concept is epigenetic inheritance. Using a contemporary biological framework based on the Extended Evolutionary Synthesis, epigenetic inheritance is shown as a missing gap despite being a key building block in modern interpretations of how evolution occurs. Epigenetic mechanisms allow quick non- or partially-genetic adaptations to environmental changes. Dynamic multi-objective optimization problems represent similar circumstances to the natural world where fitness can be determined by multiple objectives (traits), and the environment is constantly changing. We ask if the advantages that epigenetic inheritance provide in the natural world are replicated in dynamic multi-objective optimization problems. Specifically, an epigenetic blocking mechanism is applied to a state-of-the-art multi-objective genetic algorithm and its performance is compared on three sets of dynamic test functions. The mechanism shows improved performance on 12 of the 16 test problems, providing initial evidence that more algorithms should explore the wealth of epigenetic mechanisms seen in the natural world. These studies are extended to a practical problem, to demonstrate the real-world applicability of such an approach.Bio: Sizhe Yuen is a final year PhD Student at the University of Southampton. His work focuses on the development of Evolutionary Algorithms using modern biological concepts from the Extended Evolutionary Synthesis. He has also worked on using machine learning to predict behavioural patterns in Dota 2 players. He graduated in 2019 from the University of St Andrews with an MSci (Hons) in Computer Science. His research is supported by the Southampton Marine and Maritime Institute. 32-G449, Patil/Kiva

August 09, 2022

Twilight: A Differentially Private Payment Channel Network

Saar Tochner
The Hebrew University
1:00P
- 2:00P

Location

Seminar Room G882 (Hewlett Room)
Add to Calendar 2022-08-09 13:00:00 2022-08-09 14:00:00 America/New_York Twilight: A Differentially Private Payment Channel Network Abstract: Payment channel networks (PCNs) provide a faster and cheaper alternative to transactions recorded on the blockchain. Clients can trustlessly establish payment channels with relays by locking coins and then send signed payments that shift coin balances over the network’s channels. Although payments are never published, anyone can track a client’s payment by monitoring changes in coin balances over the network’s channels. We present Twilight, the first PCN that provides a rigorous differential privacy guarantee to its users. Relays in Twilight run a noisy payment processing mechanism that hides the payments they carry. This mechanism increases the relay’s cost, so Twilight combats selfish relays that wish to avoid it using a trusted execution environment (TEE) that ensures they follow its protocol. The TEE does not store the channel’s state, which minimizes the trusted computing base. Crucially, Twilight ensures that even if a relay breaks the TEE’s security, it cannot break the integrity of the PCN. We analyze Twilight in terms of privacy and cost and study the trade-off between them. We implement Twilight using Intel’s SGX framework and evaluate its performance using relays deployed on two continents. We show that a route consisting of 4 relays handles 820 payments/sec. Seminar Room G882 (Hewlett Room)

September 30, 2022

Private Event

The Technological Foundations of Prosperity Symposium

9:00A
- 5:00P

Location

Kiva
Add to Calendar 2022-09-30 9:00:00 2022-09-30 17:00:00 America/New_York The Technological Foundations of Prosperity Symposium Creating a cross-disciplinary community with computer scientists and economists faces two hurdles. The first hurdle, which is common to all interdisciplinary work, is that different disciplines tend to focus on different topics, use different methods, and be physically separated – all of which deter mutual understanding and collaboration. But these factors are not determinative, many disciplines still find ways of collaborating. One might have expected this to also be true for computer science and economics, since computer science is such an important driver of productivity growth. However the historical structure of computer science – modular with levels of abstraction – is a second hurdle that has exacerbated the separation between the disciplines. This is because changes lower in the stack, e.g. Moore’s Law and hardware architecture changes, are hidden from the user behind modular interfaces (e.g. operating system). This technical abstraction can (and often is) exploited by economists who then only look at high-level attributes of technology, e.g. spending on I.T., rather than the technical details of what is actually being implemented and on what system. For decades this worked reasonably well. But, as Moore’s Law ends, engineers are beginning to break these modularities in search of more performance. This creates an urgent need for economists and business scholars to re-engage with the technical details in order to understand how I.T. will drive prosperity. This conference will help to create this community and articulate a research agenda for it.

October 01, 2022

Private Event

The Technological Foundations of Prosperity Symposium

9:00A
- 5:00P

Location

Kiva
Add to Calendar 2022-10-01 9:00:00 2022-10-01 17:00:00 America/New_York The Technological Foundations of Prosperity Symposium Creating a cross-disciplinary community with computer scientists and economists faces two hurdles. The first hurdle, which is common to all interdisciplinary work, is that different disciplines tend to focus on different topics, use different methods, and be physically separated – all of which deter mutual understanding and collaboration. But these factors are not determinative, many disciplines still find ways of collaborating. One might have expected this to also be true for computer science and economics, since computer science is such an important driver of productivity growth. However the historical structure of computer science – modular with levels of abstraction – is a second hurdle that has exacerbated the separation between the disciplines. This is because changes lower in the stack, e.g. Moore’s Law and hardware architecture changes, are hidden from the user behind modular interfaces (e.g. operating system). This technical abstraction can (and often is) exploited by economists who then only look at high-level attributes of technology, e.g. spending on I.T., rather than the technical details of what is actually being implemented and on what system. For decades this worked reasonably well. But, as Moore’s Law ends, engineers are beginning to break these modularities in search of more performance. This creates an urgent need for economists and business scholars to re-engage with the technical details in order to understand how I.T. will drive prosperity. This conference will help to create this community and articulate a research agenda for it.