Learning Probabilistic Models for Safe Predictive Control in Unknown Environments
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Researchers rely increasingly on tools from machine learning to improve the performance of control algorithms on real world tasks and enable robots to operate for long periods of time without intervention. Many of these algorithms require a model for the dynamics of the robot. In particular, researchers designing methods for safe learning control often rely on an upper bound on model error to make guarantees about the worst-case closed-loop performance of their algorithm. There are different options for how to learn such a model of the robot dynamics. We study probabilistic models for use in the context of stochastic model predictive control. Two popular choices for learning the robot dynamics are Gaussian Process (GP) regression and various forms of local linear regression. In this paper, we present a study comparing GPs with a particular form of local linear regression for learning robot dynamics with the aim of guaranteeing safety when a robot operates in novel conditions. We show results based on experimental data from a 900 kg ground robot using vision for localisation.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it