Ethan Callanan

Ethan Callanan, Rebecca De Venezia, Victoria Armstrong et al.

For over three decades, the planning community has explored countless methods for data-driven model acquisition. These range in sophistication (e.g., simple set operations to full-blown reformulations), methodology (e.g., logic-based vs. planing-based), and assumptions (e.g., fully vs. partially observable). With no fewer than 43 publications in the space, it can be overwhelming to understand what approach could or should be applied in a new setting. We present a holistic characterization of the action model acquisition space and further introduce a unifying framework for automated action model acquisition. We have re-implemented some of the landmark approaches in the area, and our characterization of all the techniques offers deep insight into the research opportunities that remain; i.e., those settings where no technique is capable of solving.

Ethan Callanan, Amarachi Mbakwe, Antony Papadimitriou et al.

Large Language Models (LLMs) have demonstrated remarkable performance on a wide range of Natural Language Processing (NLP) tasks, often matching or even beating state-of-the-art task-specific models. This study aims at assessing the financial reasoning capabilities of LLMs. We leverage mock exam questions of the Chartered Financial Analyst (CFA) Program to conduct a comprehensive evaluation of ChatGPT and GPT-4 in financial analysis, considering Zero-Shot (ZS), Chain-of-Thought (CoT), and Few-Shot (FS) scenarios. We present an in-depth analysis of the models' performance and limitations, and estimate whether they would have a chance at passing the CFA exams. Finally, we outline insights into potential strategies and improvements to enhance the applicability of LLMs in finance. In this perspective, we hope this work paves the way for future studies to continue enhancing LLMs for financial reasoning through rigorous evaluation.

Xianzhi Li, Ethan Callanan, Abdellah Ghassel et al.

Test-time compute methods can significantly improve the reasoning capabilities and problem-solving accuracy of large language models (LLMs). However, these approaches require substantially more computational resources, with most compute wasted on exploring low-diversity branches where the model already exhibits high confidence. We observe that a small subset of uncertain reasoning steps has a disproportionately large impact on final prediction accuracy, and branching at these critical junctures tends to yield more diverse and higher-quality candidate reasoning steps. We propose Entropy-Gated Branching (EGB), which branches only at high-uncertainty steps and prunes expansions with a lightweight verifier. On mathematical and financial reasoning benchmarks, EGB improves accuracy by 22.6% over standard inference while operating 31%-75% faster across math benchmarks than test-time beam search with higher performance. Our results show that dynamic resource allocation during inference can substantially improve both efficiency and effectiveness, offering a more scalable pathway to enhanced LLM reasoning capabilities.