Behrooz Omidvar-Tehrani

Linbo Liu, Xinle Liu, Qiang Zhou et al. · amazon-science

With the rapid advancement of powerful large language models (LLMs) in recent years, a wide range of software engineering tasks can now be addressed using LLMs, significantly enhancing productivity and scalability. Numerous benchmark datasets have been developed to evaluate the coding capabilities of these models, while they primarily focus on code generation and issue-resolution tasks. In contrast, we introduce a new coding benchmark MigrationBench with a distinct focus: code migration. MigrationBench aims to serve as a comprehensive benchmark for migration from Java 8 to the latest long-term support (LTS) versions (Java 17, 21), including a full dataset and its subset selected with 5,102 and 300 repositories respectively. selected is a representative subset curated for complexity and difficulty, offering a versatile resource to support research in the field of code migration. Additionally, we provide a comprehensive evaluation framework to facilitate rigorous and standardized assessment of LLMs on this challenging task. We further propose an agentic framework and demonstrate that LLMs can effectively tackle repository-level code migration to Java 17. For the selected subset with Claude-4.5-Sonnet, our agentic framework achieves 71.67% and 53.33% success rate (pass@1) for minimal and maximal migration respectively. The dataset and evaluation source code are available at: https://huggingface.co/collections/AmazonScience/migrationbench and https://github.com/amazon-science/MigrationBench respectively.

Chanakya Ekbote, Vijay Lingam, Behrooz Omidvar-Tehrani et al.

Reinforcement Learning with Verifiable Rewards (RLVR) has emerged as a powerful framework for enhancing the reasoning capabilities of large language models (LLMs). However, existing approaches such as Group Relative Policy Optimization (GRPO) and its variants, while effective on reasoning benchmarks, struggle with agentic tasks that require iterative decision-making. We introduce Murphy, a multi-turn reflective optimization framework that extends GRPO by incorporating iterative self-correction during training. By leveraging both quantitative and qualitative execution feedback, Murphy enables models to progressively refine their reasoning across multiple turns. Evaluations on code generation benchmarks with model families such as Qwen and OLMo show that Murphy consistently improves performance, achieving up to a 8% relative gain in pass@1 over GRPO, on similar compute budgets.

Gauthier Guinet, Behrooz Omidvar-Tehrani, Anoop Deoras et al. · amazon-science

We propose a new method to measure the task-specific accuracy of Retrieval-Augmented Large Language Models (RAG). Evaluation is performed by scoring the RAG on an automatically-generated synthetic exam composed of multiple choice questions based on the corpus of documents associated with the task. Our method is an automated, cost-efficient, interpretable, and robust strategy to select the optimal components for a RAG system. We leverage Item Response Theory (IRT) to estimate the quality of an exam and its informativeness on task-specific accuracy. IRT also provides a natural way to iteratively improve the exam by eliminating the exam questions that are not sufficiently informative about a model's ability. We demonstrate our approach on four new open-ended Question-Answering tasks based on Arxiv abstracts, StackExchange questions, AWS DevOps troubleshooting guides, and SEC filings. In addition, our experiments reveal more general insights into factors impacting RAG performance like size, retrieval mechanism, prompting and fine-tuning. Most notably, our findings show that choosing the right retrieval algorithms often leads to bigger performance gains than simply using a larger language model.

Behrooz Omidvar-Tehrani

In this paper, we propose a geospatial data management framework called IRIDEF which captures and analyzes user's exploratory feedback for an enriched guidance mechanism in the context of interactive analysis. We discuss that exploratory feedback can be a proxy for decision-making feedback when the latter is scarce or unavailable. IRIDEF identifies regions of interest (ROIs) via exploratory feedback and highlights a few interesting and out-of-sight POIs in each ROI. These highlights enable the user to shape up his/her future interactions with the system. We detail the components of our proposed framework in the form of a data analysis pipeline and present the aspects of efficiency and effectiveness for each component. We also discuss evaluation plans and future directions for IRIDEF.

Sobhan Moosavi, Behrooz Omidvar-Tehrani, R. Bruce Craig et al.

Because of the increasing availability of spatiotemporal data, a variety of data-analytic applications have become possible. Characterizing driving context, where context may be thought of as a combination of location and time, is a new challenging application. An example of such a characterization is finding the correlation between driving behavior and traffic conditions. This contextual information enables analysts to validate observation-based hypotheses about the driving of an individual. In this paper, we present DriveContext, a novel framework to find the characteristics of a context, by extracting significant driving patterns (e.g., a slow-down), and then identifying the set of potential causes behind patterns (e.g., traffic congestion). Our experimental results confirm the feasibility of the framework in identifying meaningful driving patterns, with improvements in comparison with the state-of-the-art. We also demonstrate how the framework derives interesting characteristics for different contexts, through real-world examples.