Feb. 13, 2024, 5:41 a.m. | Debmalya Mandal Andi Nika Parameswaran Kamalaruban Adish Singla Goran Radanovi\'c

cs.LG updates on arXiv.org arxiv.org

We study data corruption robustness for reinforcement learning with human feedback (RLHF) in an offline setting. Given an offline dataset of pairs of trajectories along with feedback about human preferences, an $\varepsilon$-fraction of the pairs is corrupted (e.g., feedback flipped or trajectory features manipulated), capturing an adversarial attack or noisy human preferences. We aim to design algorithms that identify a near-optimal policy from the corrupted data, with provable guarantees. Existing theoretical works have separately studied the settings of corruption robust …

adversarial corruption cs.ai cs.lg data dataset features feedback human human feedback offline reinforcement reinforcement learning rlhf robust robustness study trajectory

Research Scholar (Technical Research)

@ Centre for the Governance of AI | Hybrid; Oxford, UK

HPC Engineer (x/f/m) - DACH

@ Meshcapade GmbH | Remote, Germany

Senior Analytics Engineer (Retail)

@ Lightspeed Commerce | Toronto, Ontario, Canada

Data Scientist II, BIA GPS India Operations

@ Bristol Myers Squibb | Hyderabad

Analytics Engineer

@ Bestpass | Remote

Senior Analyst - Data Management

@ Marsh McLennan | Mumbai - Hiranandani