About

My research group focuses on theoretical problems in deep learning and visual perception in robotics. We conduct multi-disciplinary research using ideas from information theory, neuroscience, and physics, in addition to classical perspectives from learning theory and computer vision. Our group is diverse, including engineers, computer scientists, mathematicians, and physicists.

Research topics

• Computer Science
• Artificial Intelligence
• Machine Learning
• Mathematics
• Geography
• Cartography
• Remote sensing
• Cognitive psychology
• Engineering
• Clinical psychology
• Psychology
• Psychiatry
• Computer vision

Selected publications

• Combinatorial optimization with Kerr solitons

  Science Advances · 2026-05-22

  articleOpen access

  The challenge of scaling digital computing motivates innovation, especially through the evolution of physical systems that mimic neural networks and combinatorial optimization problems. Light is a hyperefficient information carrier, and if efficient interactions with it could be uncovered, then direct information processing would become far more feasible. We harness an ensemble of hundreds of Kerr microresonator solitons and implement an analog feedback network to create an Ising machine with fully programmable all-to-all interactions. By increasing the feedback for self, on-diagonal interactions, each soliton exhibits a universal spin-like bifurcation, and using this palette of interactions, we solve the canonical Boolean satisfiability problem (SAT). The combination of uniform soliton interactions and the compatibility of our Ising machine with high-speed data interconnects enables rapid and precise solutions of complex SAT problems. The well-established theoretical properties of Kerr solitons bound the trade-off of optical power and time use by the machine at ~0.15 milliwatts per soliton and 1 microsecond for a single feedback step. We performed >10,000 trials on more than 100 randomly generated SAT instances to evaluate the Ising machine, demonstrating the potential to exceed the performance of benchmark digital SAT solvers. Our work highlights the convergence of optical nonlinearity, ultralow loss photonics, and optoelectronic circuits for computation-acceleration tasks.

  PublisherDOI

• A simple model of co-emergence of grid and place fields

  arXiv (Cornell University) · 2026-05-20

  preprintOpen access

  Grid cells in the medial entorhinal cortex and place cells in the hippocampus together support spatial navigation. The two regions are reciprocally connected, and there is a chicken-and-egg problem for how both arise and reinforce each other during development. Current computational accounts either derive one type from the other or use network dynamics to model the emergence of one type in isolation. We introduce a unified recurrent network model that instantiates Dale's Law (every neuron is either excitatory or inhibitory), and is trained to predict the next sensory observation from masked previous sensory observations and egocentric motion. To our knowledge, this is the first single-objective model in which grid and place cells co-emerge without supervision of either type, or reliance on pre-existing spatial-cell representations. The two kinds of spatial codes coexist across 1,000 different training configurations, with their balance set by the amount of sensory noise and masking. Without retraining, the network qualitatively reproduces experimentally observed grid fragmentation in hairpin mazes, grid merging after wall removal, lattice alignment across connected rooms, locally ordered 3D fields observed in freely flying bats, as well as the developmental order in which place cells precede grid cells. We interpret these results in terms of two complementary encoding pressures within a single sensory-prediction objective: (1) correcting errors or reconstructing missing components of sensory observations, and (2) prediction of the next sensory state during navigation. Our results suggest a circuit-level account of the co-emergence of grid and place cells, and experimentally testable predictions for the two kinds of spatial codes.

  PublisherDOI

• Code for "An Analytical Characterization of Sloppiness in Neural Networks: Insights from Linear Models"

  Zenodo (CERN European Organization for Nuclear Research) · 2026-01-06 · 1 citations

  otherOpen accessSenior author

  Code for the paper "An Analytical Characterization of Sloppiness in Neural Networks: Insights from Linear Models".

  PublisherDOI

• A simple model of co-emergence of grid and place fields

  ArXiv.org · 2026-05-20

  articleOpen access

  Grid cells in the medial entorhinal cortex and place cells in the hippocampus together support spatial navigation. The two regions are reciprocally connected, and there is a chicken-and-egg problem for how both arise and reinforce each other during development. Current computational accounts either derive one type from the other or use network dynamics to model the emergence of one type in isolation. We introduce a unified recurrent network model that instantiates Dale's Law (every neuron is either excitatory or inhibitory), and is trained to predict the next sensory observation from masked previous sensory observations and egocentric motion. To our knowledge, this is the first single-objective model in which grid and place cells co-emerge without supervision of either type, or reliance on pre-existing spatial-cell representations. The two kinds of spatial codes coexist across 1,000 different training configurations, with their balance set by the amount of sensory noise and masking. Without retraining, the network qualitatively reproduces experimentally observed grid fragmentation in hairpin mazes, grid merging after wall removal, lattice alignment across connected rooms, locally ordered 3D fields observed in freely flying bats, as well as the developmental order in which place cells precede grid cells. We interpret these results in terms of two complementary encoding pressures within a single sensory-prediction objective: (1) correcting errors or reconstructing missing components of sensory observations, and (2) prediction of the next sensory state during navigation. Our results suggest a circuit-level account of the co-emergence of grid and place cells, and experimentally testable predictions for the two kinds of spatial codes.

  PublisherOA PDF

• A simple model of the co-emergence of grid and place fields

  bioRxiv (Cold Spring Harbor Laboratory) · 2026-05-22

  articleOpen access

  Grid cells in the medial entorhinal cortex and place cells in the hippocampus together support spatial navigation. The two regions are reciprocally connected, and there is a chicken-and-egg problem for how both arise and reinforce each other during development. Current computational accounts either derive one type from the other or use network dynamics to model the emergence of one type in isolation. We introduce a unified recurrent network model that instantiates Dale's Law (every neuron is either excitatory or inhibitory), and is trained to predict the next sensory observation from masked previous sensory observations and egocentric motion. To our knowledge, this is the first single-objective model in which grid and place cells co-emerge without supervision of either type, or reliance on pre-existing spatial-cell representations. The two kinds of spatial codes coexist across 1,000 different training configurations, with their balance set by the amount of sensory noise and masking. Without retraining, the network qualitatively reproduces experimentally observed grid fragmentation in hairpin mazes, grid merging after wall removal, lattice alignment across connected rooms, locally ordered 3D fields observed in freely flying bats, as well as the developmental order in which place cells precede grid cells. We interpret these results in terms of two complementary encoding pressures within a single sensory-prediction objective: (1) correcting errors or reconstructing missing components of sensory observations, and (2) prediction of the next sensory state during navigation. Our results suggest a circuit-level account of the co-emergence of grid and place cells, and experimentally testable predictions for the two kinds of spatial codes.

  PublisherDOI

• Code for "An Analytical Characterization of Sloppiness in Neural Networks: Insights from Linear Models"

  Zenodo (CERN European Organization for Nuclear Research) · 2026-01-06

  otherOpen accessSenior author

  Code for the paper "An Analytical Characterization of Sloppiness in Neural Networks: Insights from Linear Models".

  PublisherDOI

• How Occam’s razor guides human decision-making

  eLife · 2025-12-09

  articleOpen access

  Occam’s razor is the principle that, all else being equal, simpler explanations should be preferred over more complex ones. This principle is thought to guide human decision-making, but the nature of this guidance is not known. Here we used preregistered behavioral experiments to show that people tend to prefer the simpler of two alternative explanations for uncertain data. These preferences match predictions of formal theories of model selection that penalize excessive flexibility. These penalties emerge when considering not just the best explanation but the integral over all possible, relevant explanations. We further show that these simplicity preferences persist in humans, but not in certain artificial neural networks, even when they are maladaptive. Our results imply that principled notions of statistical model selection, including integrating over possible, latent causes to avoid overfitting to noisy observations, may play a central role in human decision-making.

  PublisherDOI

• A Kerr soliton Ising machine for combinatorial optimization problems

  ArXiv.org · 2025-08-01 · 1 citations

  preprintOpen access

  The growing challenges of scaling digital computing motivate new approaches, especially through the dynamical evolution of physical systems that mimic neural networks and combinatorial optimization problems. While light is a hyper efficient information carrier, intrinsically weak light interactions make direct information processing difficult to implement. Recently, specialized nonlinear photonics have opened new controls over light fields with extraordinary bandwidth, coherence, and the emergence of strong interactions among nonlinear eigenstates like solitons. We harness an ensemble of hundreds of Kerr-nonlinear microresonator solitons and implement an analog feedback network to create an Ising machine with fully programmable all-to-all interactions. By increasing the feedback for self, on-diagonal interactions, each soliton exhibits a universal spin-like bifurcation. Using this palette of interactions amongst the entire soliton ensemble, we encode the Ising machine to solve the benchmark Boolean satisfiability problem (SAT). The combination of uniform soliton interactions and the compatibility of our Ising machine with high-speed data interconnects enables rapid and precise solutions of complex SAT problems. Indeed, the soliton properties bound the tradeoff of optical power and time use by the machine at approximately 10 mW and 1 $μ$s for a single feedback step. We performed >10,000 trials on more than 100 randomly generated SAT instances to evaluate the Ising machine, demonstrating the potential to exceed the performance of benchmark digital SAT solvers. Our work highlights the convergence of optical nonlinearity, ultralow loss photonics, and optoelectronic circuits, which can be combined for a wide range of computation-acceleration tasks.

  PublisherOA PDFDOI

• Symskill: Symbol and Skill Co-Invention for Data-Efficient and Reactive Long-Horizon Manipulation

  ArXiv.org · 2025-10-02

  preprintOpen access

  Multi-step manipulation in dynamic environments remains challenging. Imitation learning (IL) is reactive but lacks compositional generalization, since monolithic policies do not decide which skill to reuse when scenes change. Classical task-and-motion planning (TAMP) offers compositionality, but its high planning latency prevents real-time failure recovery. We introduce SymSkill, a unified framework that jointly learns predicates, operators, and skills from unlabeled, unsegmented demonstrations, combining compositional generalization with real-time recovery. Offline, SymSkill learns symbolic abstractions and goal-oriented skills directly from demonstrations. Online, given a conjunction of learned predicates, it uses a symbolic planner to compose and reorder skills to achieve symbolic goals while recovering from failures at both the motion and symbolic levels in real time. Coupled with a compliant controller, SymSkill supports safe execution under human and environmental disturbances. In RoboCasa simulation, SymSkill executes 12 single-step tasks with 85% success and composes them into multi-step plans without additional data. On a real Franka robot, it learns from 5 minutes of play data and performs 12-step tasks from goal specifications. Code and additional analysis are available at https://sites.google.com/view/symskill.

  PublisherOA PDFDOI

• RT-GuIDE: Real-Time Gaussian Splatting for Information-Driven Exploration

  IEEE Robotics and Automation Letters · 2025-09-25 · 1 citations

  article

  We propose a framework for active mapping and exploration that leverages Gaussian splatting for constructing dense maps. Further, we develop a GPU-accelerated motion planning algorithm that can exploit the Gaussian map for real-time navigation. The Gaussian map constructed onboard the robot is optimized for both photometric and geometric quality while enabling real-time situational awareness for autonomy. We show through viewpoint selection experiments that our method yields comparable Peak Signal-to-Noise Ratio (PSNR) and similar reconstruction error to state-of-the-art approaches, while being orders of magnitude faster to compute. In closed-loop physics-based simulation and real-world experiments, our algorithm achieves better map quality (at least 0.8dB higher PSNR and more than 16% higher geometric reconstruction accuracy) than maps constructed by a state-of-the-art method, enabling semantic segmentation using off-the-shelf open-set models.

  PublisherDOI

Recent grants

• CAREER: Foundations of Small Data

  NSF · $439k · 2022–2027

Frequent coauthors

• Stefano Soatto

  29 shared

• Christos Davatzikos

  University of Pennsylvania

  20 shared

• Emilio Frazzoli

  18 shared

• Rasool Fakoor

  18 shared

• Alexander J. Smola

  15 shared

• Yann LeCun

  New York University

  14 shared

• Rongguang Wang

  University of Pennsylvania

  14 shared

• Fanyang Yu

  13 shared

Education

• B.S., Aerospace Engineering

  IIT Bombay

  2010

• M.S., Aeronautics & Astronautics

  Massachusetts Institute of Technology

  2012

• Ph.D., Computer Science

  University of California, Los Angeles

  2018

Awards & honors

• NSF CAREER Award