About

Tal Cohen is an Associate Professor and Postdoctoral Chair at the Massachusetts Institute of Technology in the Department of Civil and Environmental Engineering, specializing in Sustainable Materials & Infrastructure. His research interests include nonlinear solid mechanics, material growth, and material instabilities. Cohen's work involves understanding complex behaviors of materials under various conditions, with applications spanning biological materials, growth processes, and material failure mechanisms. He holds a Ph.D. from the Faculty of Aerospace Engineering at the Technion in Israel, earned in 2014, along with a master's and bachelor's degree from the same institution. Cohen has received numerous awards, including the MIT Arthur C. Smith Award in 2024, the Eshelby Mechanics Award for Young Faculty in 2023, and the NSF CAREER Award in 2020. His contributions to the field are recognized through his research publications and his role in advancing knowledge in mechanics, growth, and material stability.

Research topics

• Materials science
• Composite material
• Mechanics
• Cell biology
• Chemical physics
• Thermodynamics
• Biomedical engineering
• Genetics
• Nanotechnology
• Chemistry
• Statistical physics
• Medicine
• Physics
• Biology
• Microbiology

Selected publications

• Constraint Architecture of Physical AI Deployment: A Coupled Interaction Model

  Zenodo (CERN European Organization for Nuclear Research) · 2026-04-20

  preprintOpen access1st authorCorresponding

  Physical AI deployment – autonomous systems that perceive, decide, and act in the physical world underreal consequence – is failing at the pilot-to-production boundary at non-trivial rates across multipleindustrial settings. The paper’s central contribution is the Constraint Interaction Model (CIM), which demonstrates that the four laws form a fully coupled system: every law interacts with every other, producing six pairwiseinteractions that are shown to be directionally asymmetric in four of the six cases, yielding up to twelvedistinct diagnostic signatures across the four-law space.

  PublisherDOI

• Constraint Architecture of Physical AI Deployment: A Coupled Interaction Model

  Zenodo (CERN European Organization for Nuclear Research) · 2026-04-20

  preprintOpen access1st authorCorresponding

  Physical AI deployment – autonomous systems that perceive, decide, and act in the physical world underreal consequence – is failing at the pilot-to-production boundary at non-trivial rates across multipleindustrial settings. The paper’s central contribution is the Constraint Interaction Model (CIM), which demonstrates that the four laws form a fully coupled system: every law interacts with every other, producing six pairwiseinteractions that are shown to be directionally asymmetric in four of the six cases, yielding up to twelvedistinct diagnostic signatures across the four-law space.

  PublisherDOI

• Constraint Architecture of Physical AI Deployment: A Coupled Interaction Model

  Zenodo (CERN European Organization for Nuclear Research) · 2026-04-20

  preprintOpen access1st authorCorresponding

  Physical AI deployment – autonomous systems that perceive, decide, and act in the physical world underreal consequence – is failing at the pilot-to-production boundary at non-trivial rates across multipleindustrial settings. The paper’s central contribution is the Constraint Interaction Model (CIM), which demonstrates that the four laws form a fully coupled system: every law interacts with every other, producing six pairwiseinteractions that are shown to be directionally asymmetric in four of the six cases, yielding up to twelvedistinct diagnostic signatures across the four-law space.

  PublisherDOI

• Habitat as the Missing Layer in Physical AI Deployment

  Zenodo (CERN European Organization for Nuclear Research) · 2026-04-20

  preprintOpen access1st authorCorresponding

  Physical AI deployment, autonomous systems that perceive, decide, and act in the physical world under real consequence, is failing at the pilot-to-production boundary at non-trivial rates across multiple industrial settings. One underlying reason this paper examines is that existing integration disciplines, developed over decades for deterministic software and traditional industrial systems, are structurally under-fitted to the distinct demands of probabilistic, learning-based agents operating where errors are irreversible and institutional acceptance gates scaling.

  PublisherDOI

• Habitat as the Missing Layer in Physical AI Deployment

  Zenodo (CERN European Organization for Nuclear Research) · 2026-04-20

  preprintOpen access1st authorCorresponding

  Physical AI deployment, autonomous systems that perceive, decide, and act in the physical world under real consequence, is failing at the pilot-to-production boundary at non-trivial rates across multiple industrial settings. One underlying reason this paper examines is that existing integration disciplines, developed over decades for deterministic software and traditional industrial systems, are structurally under-fitted to the distinct demands of probabilistic, learning-based agents operating where errors are irreversible and institutional acceptance gates scaling.

  PublisherDOI

• A unified thermo-chemo-mechanical framework for bulk and frontal polymerization: Coupled kinetics and front stability

  Journal of the Mechanics and Physics of Solids · 2026-05-15

  articleSenior authorCorresponding
  PublisherDOI

• Challenging common notions on how eggs break and the role of strength versus toughness

  Communications Physics · 2025-05-08 · 3 citations

  articleOpen accessSenior author

  One experiment commonly used to teach young students about the response of structures to dynamic loading is the “egg drop challenge”, in which students design a device to protect an egg from cracking after a fall from a specified height. Relevant to this activity is the choice of orientation of the egg to decrease the probability of fracture. In this study, we contest the commonly held belief that an egg is strongest when dropped vertically on its end. Through hundreds of experiments and a set of static and dynamic simulations, we demonstrate a statistically significant decrease in the likelihood that an egg breaks when oriented horizontally as opposed to vertically, and offer a concrete and intuitive explanation as to why this is the case. These results and the associated analysis demonstrate the importance of specificity of language and the dangers of appealing to “common sense” in the physics classroom while having wide-ranging implications due to the ubiquity of shell structures in nature and in the man-made world. Common assumption dictates that an egg exhibits greater structural resistance when dropped on its end rather than on its side. To test this supposition, the authors perform static and dynamic loading tests on hundreds of eggs, supported with finite element simulations. Contrary to expectations, the results indicate that vertical orientation may in fact be the weaker of the two axes.

  PublisherOA PDFDOI

• An accessible instrument for measuring soft material mechanical properties

  Review of Scientific Instruments · 2025-04-01 · 1 citations

  articleOpen accessSenior author

  Soft material research has seen significant growth in recent years, with emerging applications in robotics, electronics, and healthcare diagnostics where understanding the material mechanical response is crucial for precision design. Traditional methods for measuring nonlinear mechanical properties of soft materials require specially sized samples that are extracted from their natural environment to be mounted on the testing instrument. This has been shown to compromise data accuracy and precision in various soft and biological materials. To overcome this, the Volume Controlled Cavity Expansion (VCCE) method was developed. This technique tests soft materials by controlling the formation rate of a liquid cavity inside the materials at the tip of an injection needle and simultaneously measuring the resisting pressure that describes the material response. Despite VCCE's early successes, expansion of its application beyond academia has been hindered by cost, size, and expertise. In response to this, the first portable, benchtop instrument utilizing VCCE is presented here. This device, built with affordable, readily available components and open-source software, streamlines VCCE experimentation without sacrificing performance or precision. It is especially suitable for space-limited settings and designed for use by non-experts, promoting widespread adoption. The instrument's efficacy was demonstrated through testing polydimethylsiloxane samples of varying stiffness. This study not only validates instrument performance but also sets the stage for further advancements and broader applications in soft material testing. All data, along with acquisition, control, and post-processing scripts, are made available on GitHub.

  PublisherDOI

• Lifting generators in connected Lie groups

  Journal of Algebra · 2025-10-10

  article1st authorCorresponding
  PublisherDOI

• Interfacial cavitation with surface tension: New insights into failure of particle reinforced polymers

  Journal of the Mechanics and Physics of Solids · 2025-10-08

  articleSenior authorCorresponding
  PublisherDOI

Recent grants

• CAREER: Nonlinear Dynamics of Shear Wave Propagation and Shock Formation in Soft and Biological Materials

  NSF · $645k · 2020–2026

Frequent coauthors

• S. Chockalingam

  American Institute of Aeronautics and Astronautics

  15 shared

• David Durban

  Technion – Israel Institute of Technology

  13 shared

• Mrityunjay Kothari

  Providence College

  13 shared

• Jing Yan

  Yale University

  12 shared

• Jian Li

  China Academy of Railway Sciences

  11 shared

• Thomas Henzel

  Massachusetts Institute of Technology

  10 shared

• Hannah Varner

  9 shared

• Chockalingam Senthilnathan

  8 shared

Labs

• Tal CohenPI

Education

• Ph.D., Civil Engineering

  Massachusetts Institute of Technology

  2006

• M.S., Civil Engineering

  Massachusetts Institute of Technology

  2002

• B.S., Civil Engineering

  Technion - Israel Institute of Technology

  2000

Awards & honors

• MIT Arthur C. Smith Award, 2024
• Eshelby Mechanics Award for Young Faculty, 2023
• NSF CAREER Award, 2020
• ONR Young Investigator Award, 2020
• ARO Young Investigator Award, 2019