Tag: MIT research

  • Latest MIT Research in Artificial Intelligence: From Fake News Detection to AI Chemistry Models

    Latest MIT Research in Artificial Intelligence: From Fake News Detection to AI Chemistry Models

    MIT News: Artificial Intelligence

    Massachusetts Institute of Technology continues to lead cutting-edge research in artificial intelligence, with recent breakthroughs spanning fake news detection, AI ethics, workforce impacts, and chemistry modeling. Below is a roundup of the latest developments from MIT’s AI and computing labs.

    The Consequences of Relying on AI for Accurate News

    A Media Lab study reveals that dependence on AI, similar to how GPS has weakened navigation skills, can reduce our ability to detect fake news. The findings underscore the importance of human oversight in information verification.

    The Crucial Human Component in Computing and AI

    The MIT Ethics of Computing Research Symposium brought together experts to discuss ethical and social impacts of technology, emphasizing that human judgment remains essential in AI development.

    PATH Initiative: Boosting AI Training for Industry Jobs

    MIT RAISE and Georgia State University announced a collaborative initiative to connect universities, community colleges, industry, and government, expanding industry-aligned AI training and career pathways.

    NSF Renews Support for AI and Physics Institute

    The Institute for Artificial Intelligence and Fundamental Interactions (IAIFI) enters its second phase with increased funding, broader ambitions, and a growing community at the frontier of AI and fundamental physics.

    Teaching AI Agents to Ask Better Questions with ‘Battleship’

    MIT researchers used the classic game Battleship as a test bed for AI agents, discovering a small AI model can outperform the biggest ones at just 1% of the cost, demonstrating efficiency gains in AI reasoning.

    Tod Machover Receives George Peabody Medal

    The renowned composer and technologist was honored for his contributions to music and technology, receiving the highest distinction from the Peabody Institute.

    MIT Researchers Teach AI Models to Interpret Charts

    The new ChartNet training dataset aims to improve the accuracy of vision-language models, enabling better analysis of business trends and scientific figures.

    Media Advisory: MIT to Establish Regional Quantum Hub

    With a $25 million investment from Massachusetts, MIT will build a new shared-use facility serving as a statewide quantum toolbox, advancing quantum computing and AI integration.

    Will AI Create Jobs for Young, Skilled Workers?

    A study of postwar U.S. labor trends shows which workers historically filled new tech-enabled jobs, offering insights into AI’s potential impact on employment.

    Building AI Models That Understand Chemical Principles

    Connor Coley works at the intersection of chemistry and machine learning to discover and design new drug compounds, showcasing AI’s potential in scientific discovery.

    Justin Solomon Appointed Associate Dean of Engineering Education

    The MIT electrical engineering and computer science faculty member will focus on innovation in engineering education and new pedagogical approaches.

    Two from MIT Named 2026 Knight-Hennessy Scholars

    The prestigious fellowship funds graduate studies at Stanford University, recognizing MIT’s talent in AI and related fields.

    Expanding MIT’s Global Reach Through Universal Learning

    Dimitris Bertsimas and Megan Mitchell discuss the motivation behind Universal Learning, a new MIT Open Learning educational initiative that leverages AI for accessible education worldwide.

    Universal AI: A Pathway to AI Fluency for Everyone

    New AI education program from MIT Open Learning debuts with AI-powered personalization and a free introductory course, making AI accessible to learners everywhere.

    Study: Firms Often Use Automation to Control Wages

    MIT economists found that companies tend to target employees earning a ‘wage premium’ with automation, increasing inequality but not necessarily productivity.

    Source: Massachusetts Institute of Technology. Visit MIT News for the full articles and more.

  • MIT Study Reveals Why Trio Comparisons Outperform Pairwise for Predicting Preferences

    MIT Study Reveals Why Trio Comparisons Outperform Pairwise for Predicting Preferences

    A new paper from MIT researchers provides a major upgrade to the nearly century-old idea of random utility models (RUMs), showing that asking people to rank three options instead of just two can reveal hidden correlations that dramatically improve preference predictions.

    In 1927, psychologist L. L. Thurstone laid the foundation for random utility models, which assume that when people choose among alternatives, they select the one with the highest subjective value, even if they cannot assign a specific number to that choice. These models are inherently random because preferences vary across individuals and even within the same person over time.

    RUMs are widely used by governments and companies to predict behavior in counterfactual scenarios, such as how commuters would react to a road closure or how to allocate a budget to maximize public good. Despite nearly a century of refinement, the standard approach relies heavily on pairwise comparisons (e.g., “Do you prefer A or B?”) because people find it easier to compare two items than to assign a numerical rating.

    However, the MIT team — Yeshwanth Cherapanamjeri, Gabriele Farina, Constantinos Daskalakis, and Sobhan Mohammadpour — proved that pairwise comparisons alone cannot capture correlations between preferences. For example, someone who favors gun control is likely also to support government-funded child care, or a fan of independent movies may also enjoy foreign films but dislike blockbusters. Ignoring these correlations leads to inaccurate models.

    The key breakthrough, presented at the International Conference on Learning Representations in Rio de Janeiro, is that correlations become detectable when large numbers of people rank three alternatives in order of preference. The same information can also be obtained from a combination of best-of-three and best-of-two choices. The researchers developed an efficient algorithm to merge individual triplet rankings into a single model that captures the full picture.

    “This paper provides a crucial breakthrough,” says Emma Frejinger, a computer scientist at the University of Montreal. “It mathematically proves why traditional data collection fails and demonstrates that simply asking users for their best-of-three choices unlocks the ability to accurately train these powerful models.”

    The work has direct implications for AI alignment. Large language models (LLMs) are often trained by having humans rank candidate outputs — a process that can be made far more effective by using triplet comparisons. As Daskalakis notes, “RUMs play a central role in the commercial viability and usefulness of LLMs.”

    The team’s findings also show that the number of experiments needed does not grow exponentially with the number of items in a catalog, making the approach practical for real-world applications like streaming services, e-commerce, and political polling.

    “This finding provides a highly practical roadmap for collecting better data to drive more accurate optimizations,” adds Frejinger.

    Looking ahead, the MIT researchers believe that building and refining utility models will remain a vibrant area of research, critical to aligning AI systems with human preferences and to sustaining the internet economy.

  • Generalists Outperform Specialists in Certain Game Theory Scenarios, MIT Study Finds

    Generalists Outperform Specialists in Certain Game Theory Scenarios, MIT Study Finds

    In a surprising twist that challenges long-held assumptions in game theory, a new MIT study shows that general-purpose algorithms called policy gradient methods can outperform specialized game-theoretic algorithms in certain imperfect-information games. The findings, presented at the International Conference on Learning Representations in Rio De Janeiro, could reshape how artificial intelligence agents are trained to make decisions in competitive, real-world scenarios.

    Imperfect-information games—where players don’t know everything about their opponents—are common in life, from poker and bidding wars to military operations and financial negotiations. For decades, the prevailing belief was that algorithms specifically designed for these games, grounded in game theory, would always outshine general-purpose alternatives. However, the MIT-led team discovered that policy gradient methods, originally developed in the 1990s for single-agent decision-making, often perform better and with greater efficiency.

    The researchers created a benchmark to fairly evaluate different algorithms, measuring performance through a concept called exploitability—how well a player does against a worst-case adversary. In experiments involving five games, including Phantom Tic-Tac-Toe, imperfect-information Hex, and Liar’s Dice, neural networks trained with policy gradient algorithms consistently achieved lower exploitability scores than those trained with game-theoretic algorithms.

    “Our study showed that policy gradient methods can work better than these specialized algorithms, and that the specialized algorithms may not work as well as people thought,” said Samuel Sokota, a co-author from Carnegie Mellon University. The team’s benchmarking software, which they have made freely available, allows others to test and compare algorithms with just a single line of code added to the OpenSpiel library.

    The implications extend far beyond board games. “Hidden information is a very important property of the world,” said Eugene Vinitsky of New York University, another co-author. “It pervades military operations, trading scenarios, and negotiations—all of which are carried out under conditions of hidden information. The idea that we can improve on these games suggests that we can also do better in these other settings as well.”

    Ian Gemp, a computer scientist and game theory expert at Google DeepMind not involved in the study, called the results encouraging: “This work serves as a compelling reminder that modernizing classical tools remains a highly productive path for solving complex strategic problems.”

  • Policy Gradient Methods Outperform Specialized Game Theory Algorithms in Imperfect-Information Games

    Policy Gradient Methods Outperform Specialized Game Theory Algorithms in Imperfect-Information Games

    A new MIT-led study challenges long-held assumptions in game theory, demonstrating that general-purpose policy gradient methods can outperform specialized game-theoretic algorithms in imperfect-information, zero-sum games. The research, presented at the International Conference on Learning Representations, provides a benchmark for evaluating algorithms that train neural networks to compete in strategic interactions where players have hidden information.

    The team, including MIT PhD student Sobhan Mohammadpour and Assistant Professor Gabriele Farina, found that policy gradient methods—originally developed in the 1990s for decision-making—achieved lower exploitability scores than game-theory-based approaches in games like Phantom Tic-Tac-Toe, imperfect-information Hex, and Liar’s Dice. Exploitability measures how well a player performs against a worst-case adversary; a score of zero indicates perfect play.

    “It had been pretty much taken for granted that specialized game-theoretic algorithms were the right approach,” said co-author Samuel Sokota. “Our study showed that policy gradient methods can work better than these specialized algorithms.” The researchers attribute the oversight to a lack of rigorous benchmarking, which they have now addressed by releasing a freely available benchmark tool that runs on ordinary laptops.

    The benchmark, built on OpenSpiel, allows researchers to train and compare algorithms on games with up to 30 billion states. Farina emphasized that the term “game” applies broadly to multi-agent strategic interactions, including military operations, trading, and negotiations—all of which involve hidden information. “The idea that we can improve on these games suggests that we can also do better in these other settings,” said co-author Eugene Vinitsky.

    Ian Gemp of Google DeepMind praised the work: “This work serves as a compelling reminder that modernizing classical tools like policy gradient methods remains a highly productive path for solving complex strategic problems.”

  • New Machine Learning Approach Accurately Predicts Metal Alloy Properties

    New Machine Learning Approach Accurately Predicts Metal Alloy Properties

    Companies in aerospace, energy, and computing constantly seek new materials to boost performance, but understanding how those materials will behave in real-world applications often requires costly and time-consuming physical testing. Traditional simulation techniques struggle to model the complex chemical arrangements found in most solid materials, particularly metal alloys. Now, MIT researchers have developed a method that accurately models the behavior of metals regardless of their chemical complexity.

    At the core of the approach are machine-learning models that simulate materials atom by atom. The researchers improved these models by building training datasets that capture the diversity of atomic environments in chemically disordered materials. Using information theory, they optimized the training sets to include a wider variety of local chemical configurations, replacing redundant examples with new ones to make the data more informative.

    In a paper published in Science Advances, the team demonstrated that their technique could accurately predict material properties for a diverse group of metal alloys under various conditions. The approach also showed promise for developing new materials, especially in scenarios where experimentation is expensive.

    “The focus is metallic alloys, but this could be adapted to other materials like semiconductors,” says senior author Rodrigo Freitas, MIT’s TDK Career Development Professor in Materials Science and Engineering. “You could use this approach to create new sustainable steels, new materials for aerospace, and more.”

    The researchers applied their technique to create machine-learning training datasets for several chemically diverse alloys. Models trained on these datasets outperformed larger models from companies like Google and Microsoft in accuracy. The team also validated their predictions against experimental data on atomic ordering in alloys.

    One key test involved predicting phase diagrams—maps showing which phases of an alloy are stable at different temperatures and compositions. The models closely matched experimental data, a crucial step for connecting materials modeling to real-world processing decisions like welding, casting, or heat-treating.

    The researchers are now using the method to study how changing an alloy’s composition affects mechanical properties and radiation tolerance, aiming to design materials that remain strong and damage-tolerant in harsh environments. They also plan to make the approach easier to integrate with existing tools used by materials engineers.

    “Industry isn’t going to change the way they do things if what you’re creating doesn’t fit into their existing operating procedures,” Freitas notes. “The goal is to make these predictions useful where materials decisions are actually made.”

    The research was supported by the U.S. Air Force Office of Scientific Research.