Updates
May '24 |
Our Random Latent Exploration paper was accepted to ICML '24! |
Mar '24 |
Awarded the Goldwater Scholarship! |
Jan '24 |
Our Energy Efficient Locomotion paper was accepted to ICRA '24! |
|
|
Random Latent Exploration for Deep Reinforcement Learning
Srinath Mahankali,
Zhang-Wei Hong,
Ayush Sekhari,
Alexander Rakhlin,
Pulkit Agrawal
ICML, 2024
project page /
paper /
abstract
We improve exploration in both discrete and continuous control domains by optimizing random reward functions parameterized by a sampled latent vector.
|
|
Maximizing Quadruped Velocity by Minimizing Energy
Srinath Mahankali*,
Chi-Chang Lee*,
Gabriel B. Margolis,
Zhang-Wei Hong,
Pulkit Agrawal
ICRA, 2024
project page /
paper
We train energy-efficient policies for quadruped locomotion tasks while improving task performance
through constrained reinforcement learning.
|
|
Does Novelty-Based Exploration Maximize Novelty?
Srinath Mahankali,
Zhang-Wei Hong,
Pulkit Agrawal
preprint, 2023
paper
Randomly generated rewards can explain a significant fraction of exploration improvements from novelty-based intrinsic motivation.
|
|
Norm-dependent convergence and stability of the inverse scattering series for diffuse and scalar waves
Srinath Mahankali,
Yunan Yang
Inverse Problems, 2023
paper /
abstract
We prove bounds on the convergence and stability of the inverse scattering series
under different Sobolev norms, finding conditions under which the radius of convergence
and stability are improved.
|
|
Randomly Initialized One-Layer Neural Networks Make Data Linearly Separable
Promit Ghosal,
Srinath Mahankali,
Yihang Sun
arXiv preprint, 2022
paper /
abstract
Randomly initialized one-layer neural networks, with high probability, make datasets linearly separable.
|
|
The convexity of optimal transport-based waveform inversion for certain structured velocity models
Srinath Mahankali
SIAM Undergraduate Research Online, 2021
paper /
abstract
Full waveform inversion with an optimal transport-based objective has superior convexity
compared to the standard least-squares objective function for certain
velocity models.
|
|