Neurosymbolic Imitation Learning with Human Guidance: A Privileged Information Approach
Nikhilesh Prabhakar*, Varun Balaji*, Athresh Karanam, Kristian Kersting, Sriraam Natarajan · 2026
Read Neurosymbolic Imitation Learning with Human Guidance: A Privileged Information Approach by Nikhilesh Prabhakar*, Varun Balaji*, Athresh Karanam, Kristian Kersting, Sriraam Natarajan (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-07166.txt.
Description
Nikhilesh Prabhakar and four co-authors combine neural and symbolic learning so imitation can use extra training-only information. Pure neural copying handles rich data but needs many examples and overfits easily. Pure symbolic methods generalize well yet struggle with high-dimensional input like images.
Their hybrid approach uses extra information only during training, here human gaze data, to guide learning. Neural perception handles the raw input, while rule structure supports spread to new cases. Evaluations reported by the authors show gains in impact, speed, and spread, making the approach attractive where expert demonstrations are scarce but extra training signals exist.
Catalog record
| CREATOR | Nikhilesh Prabhakar*, Varun Balaji*, Athresh Karanam, Kristian Kersting, Sriraam Natarajan |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 32 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.07166