arXiv subjects

William Ward

Publications and source records attributed to William Ward.

8 recordsLinked to original sources

arXiv2026

Human-Alignment, Calibration, and Activation Patterns in Large Language Model Uncertainty

Uncertainty Quantification is a large and growing subfield of large language model behavioral analysis. Primarily to recognize and combat hallucination, the field has largely focused on measuring and improving calibration, the accuracy of uncertainty judgments to task efficacy. In this work, we investigate the relatively underexplored question of how similar large language model uncertainty is to human uncertainty. We investigate the presence and strength of human-similar uncertainty signals, deemed uncertainty alignment, in large language model overt behavior and internal activation patterns. We identify whether the models show evidence of simultaneous alignment and calibration on a variety of datasets covering both multiple choice and open ended factual recall. And we characterize the effect of instruct fine-tuning on each of these facets.

cs.CL↗

arXiv2025

Zero to Autonomy in Real-Time: Online Adaptation of Dynamics in Unstructured Environments

Autonomous robots must go from zero prior knowledge to safe control within seconds to operate in unstructured environments. Abrupt terrain changes, such as a sudden transition to ice, create dynamics shifts that can destabilize planners unless the model adapts in real-time. We present a method for online adaptation that combines function encoders with recursive least squares, treating the function encoder coefficients as latent states updated from streaming odometry. This yields constant-time coefficient estimation without gradient-based inner-loop updates, enabling adaptation from only a few seconds of data. We evaluate our approach on a Van der Pol system to highlight algorithmic behavior, in a Unity simulator for high-fidelity off-road navigation, and on a Clearpath Jackal robot, including on a challenging terrain at a local ice rink. Across these settings, our method improves model accuracy and downstream planning, reducing collisions compared to static and meta-learning baselines.

cs.RO↗

arXiv2025

Online Adaptation of Terrain-Aware Dynamics for Planning in Unstructured Environments

Autonomous mobile robots operating in remote, unstructured environments must adapt to new, unpredictable terrains that can change rapidly during operation. In such scenarios, a critical challenge becomes estimating the robot's dynamics on changing terrain in order to enable reliable, accurate navigation and planning. We present a novel online adaptation approach for terrain-aware dynamics modeling and planning using function encoders. Our approach efficiently adapts to new terrains at runtime using limited online data without retraining or fine-tuning. By learning a set of neural network basis functions that span the robot dynamics on diverse terrains, we enable rapid online adaptation to new, unseen terrains and environments as a simple least-squares calculation. We demonstrate our approach for terrain adaptation in a Unity-based robotics simulator and show that the downstream controller has better empirical performance due to higher accuracy of the learned model. This leads to fewer collisions with obstacles while navigating in cluttered environments as compared to a neural ODE baseline.

cs.RO↗

arXiv2024

Joint Verification and Refinement of Language Models for Safety-Constrained Planning

Large language models possess impressive capabilities in generating programs (e.g., Python) from natural language descriptions to execute robotic tasks. However, these generated programs often contain errors that violate externally given task specifications. Without an effective method to verify their correctness, the reliable deployment of language models in real-world systems is practically infeasible. We develop a method that converts generated robot programs into an automaton-based representation and verifies them against task-relevant safety specifications. We establish a theorem that any arbitrary combination of the verified programs will also satisfy the safety specifications. Hence, the method eliminates the need to verify complex programs composed of multiple simpler ones, reducing computation complexity. We then introduce an automated fine-tuning procedure that leverages verification outcomes for supervision. By applying the theorem, this procedure only requires training the model to generate safe sub-components, thereby improving training efficiency. Empirical results on robot applications show a 30 percent increase in the probability of generating specification-compliant programs, with training time reduced by half compared to fine-tuning on generating full programs.

cs.AI↗

arXiv2024

Entanglement-Enhanced Neyman-Pearson Target Detection

Quantum illumination (QI) provides entanglement-enabled target-detection enhancement, despite operating in an entanglement-breaking environment. Existing experimental studies of QI have utilized a Bayesian approach, assuming that the target is equally likely to be present or absent before detection, to demonstrate an advantage over classical target detection. However, such a premise breaks down in practical operational scenarios in which the prior probability is unknown, thereby hindering QI's applicability to real-world target-detection scenarios. In this work, we adopt the Neyman-Pearson criterion in lieu of the error probability for equally likely target absence or presence as our figure of merit for QI. We demonstrate an unconditional quantum advantage over the optimal classical-illumination protocol as benchmarked by the receiver operating characteristic, which examines detection probability versus false-alarm probability without resorting to known prior probabilities. Our work represents a critical advancement in adapting quantum-enhanced sensing to practical operational settings.

quant-ph↗

arXiv2023

Fine-Tuning Language Models Using Formal Methods Feedback

Although pre-trained language models encode generic knowledge beneficial for planning and control, they may fail to generate appropriate control policies for domain-specific tasks. Existing fine-tuning methods use human feedback to address this limitation, however, sourcing human feedback is labor intensive and costly. We present a fully automated approach to fine-tune pre-trained language models for applications in autonomous systems, bridging the gap between generic knowledge and domain-specific requirements while reducing cost. The method synthesizes automaton-based controllers from pre-trained models guided by natural language task descriptions. These controllers are verifiable against independently provided specifications within a world model, which can be abstract or obtained from a high-fidelity simulator. Controllers with high compliance with the desired specifications receive higher ranks, guiding the iterative fine-tuning process. We provide quantitative evidences, primarily in autonomous driving, to demonstrate the method's effectiveness across multiple tasks. The results indicate an improvement in percentage of specifications satisfied by the controller from 60% to 90%.

cs.AI↗

arXiv2023

Active Inverse Learning in Stackelberg Trajectory Games

Game-theoretic inverse learning is the problem of inferring a player's objectives from their actions. We formulate an inverse learning problem in a Stackelberg game between a leader and a follower, where each player's action is the trajectory of a dynamical system. We propose an active inverse learning method for the leader to infer which hypothesis among a finite set of candidates best describes the follower's objective function. Instead of using passively observed trajectories like existing methods, we actively maximize the differences in the follower's trajectories under different hypotheses by optimizing the leader's control inputs. Compared with uniformly random inputs, the optimized inputs accelerate the convergence of the estimated probability of different hypotheses conditioned on the follower's trajectory. We demonstrate the proposed method in a receding-horizon repeated trajectory game and simulate the results using virtual TurtleBots in Gazebo.

cs.GT↗

arXiv2011

Astronomical Sky Quality Near Eureka, in the Canadian High Arctic

Nighttime visible-light sky brightness and transparency are reported for the Polar Environment Research Laboratory (PEARL), located on a 610-m high ridge near the Eureka research station, on Ellesmere Island, Canada. Photometry of Polaris obtained in V band with the PEARL All Sky Imager (PASI) over two winters is supported by standard meteorological measurements and visual estimates of sky conditions from sea level. These data show that during the period of the study, October through March of 2008/09 and 2009/10, the sky near zenith had a mean surface brightness of 19.7 mag/square-arcsec when the sun was more than 12 deg below the horizon, reaching 20.7 mag/square-arcsec during astronomical darkness with no moon. Skies were without thick cloud and potentially usable for astronomy 86% of the time (extinction <2 mag). Up to 68% of the time was spectroscopic (<0.5 mag), attenuated by ice crystals, or clear with stable atmospheric transparency. Those conditions can persist for over 100 hours at a time. Further analysis suggests the sky was entirely free of ice crystals (truly photometric) 48+/-3% of the time at PEARL in winter, and that a higher elevation location nearby may be better.

astro-ph.IM↗