arXiv subjects

Pavan C Shekar

Publications and source records attributed to Pavan C Shekar.

5 recordsLinked to original sources

arXiv2026

GitOfThoughts: Version-Controlled Reasoning and Agent Memory You Can Replay, Diff, and Merge

Large language model reasoning leaves no trace once it is done. The steps of a chain of thought disappear when the context window closes, a pruned search branch is just gone, and memory buffers cannot be diffed, merged, or audited. Code, infrastructure, and experiments are all version-controlled. Reasoning is not. GitOfThoughts stores an agent's reasoning tree as a git repository. Every scored thought becomes a commit, scores become notes, outcomes become tags, and retrieval is just git log over the agent's own history. We use this to test something simple. Does giving an agent memory from past problems actually make it more accurate? We tried five memory stores (none, a markdown file, a vector database, a graph, and git) across two benchmarks, two model sizes, and several pre-registered repeat experiments. The answer, on new problems, is no, including one promising early result that did not hold up when we repeated it. Memory only helps once the problem being solved is nearly identical to something already in memory (cosine similarity above about 0.8); below that, it does nothing. In other words, the model is finding the answer rather than learning the method. Even a model 4.5x larger still cannot pull a reusable method out of a worked example; it just gets better at spotting near-copies. The only thing that reliably helped on new problems was generating several answers and picking the most common one (self-consistency). So the case for using git as the memory store is not that it retrieves better. It is that it gives auditability, history, and the ability to merge two agents' memories, at no cost to accuracy.

cs.AI↗

arXiv2026

ReTreVal: Reasoning Tree with Validation and Cross-Problem Memory for Large Language Models

Every existing inference-time reasoning framework discards all failure context at problem boundaries, leaving a model solving problem 500 no wiser than it was on problem 1. We present ReTreVal (Reasoning Tree with Validation), a training-free framework that closes this gap through adaptive tree exploration with tool-augmented node refinement, typed-failure backtracking that injects categorized error context into the recovered branch, and a self-rewriting memory that accumulates and revises strategy entries across problems, enabling inference-time cross-problem learning on any fixed, unmodified LLM without fine-tuning. ReTreVal achieves 85.8% pass@1 on MATH-500 (+8.6 pp over Zero-Shot CoT, +8.6 pp over the strongest baseline Self-Refine) and 54.4% on MMLU-Pro (+15.3 pp over Self-Refine), with a 3.4:1 win-to-regression ratio confirming genuine error recovery rather than noise. These capabilities, previously requiring gradient updates, allow a 32B model to compete with much larger single-pass systems.

cs.AI↗

arXiv2025

Adaptive Minds: Empowering Agents with LoRA-as-Tools

We investigate a framework in which LoRA adapters are treated as callable tools that a base language model can dynamically select and invoke. We hypothesize that, when adapters are trained to provide strong domain-specific gains and are exposed with clear metadata, a base model can reliably route queries to the appropriate expert, effectively aggregating the benefits of many specialized adapters within a single framework. We introduce Adaptive Minds, a general framework within which we study both single-step routing and multi-step agentic reasoning. In this setting, the agent can iteratively invoke multiple adapters alongside other tools (e.g., external APIs, retrieval systems, or execution environments) and reason over their outputs across multiple steps. This reframes adapters as modular skills or memory units that can be composed during reasoning rather than statically applied. In our evaluation, the routing layer reaches 98.3% accuracy on a 30-adapter library, and well-trained specialists provide +4.6 to +84.0 percentage points of strict-scorer gain across nine task families under a single shared training recipe; the AM router aggregates these gains within 5 pp of the direct specialist on every benchmark whose queries surface domain signal. Our findings suggest that the effectiveness of this approach depends on the quality and specialization of individual adapters, and that enabling flexible composition of many such experts can significantly expand the practical capabilities of language model agents, moving toward more general, tool-augmented intelligence.

cs.AI↗

arXiv2025

HyperFake: Hyperspectral Reconstruction and Attention-Guided Analysis for Advanced Deepfake Detection

Deepfakes pose a significant threat to digital media security, with current detection methods struggling to generalize across different manipulation techniques and datasets. While recent approaches combine CNN-based architectures with Vision Transformers or leverage multi-modal learning, they remain limited by the inherent constraints of RGB data. We introduce HyperFake, a novel deepfake detection pipeline that reconstructs 31-channel hyperspectral data from standard RGB videos, revealing hidden manipulation traces invisible to conventional methods. Using an improved MST++ architecture, HyperFake enhances hyperspectral reconstruction, while a spectral attention mechanism selects the most critical spectral features for deepfake detection. The refined spectral data is then processed by an EfficientNet-based classifier optimized for spectral analysis, enabling more accurate and generalizable detection across different deepfake styles and datasets, all without the need for expensive hyperspectral cameras. To the best of our knowledge, this is the first approach to leverage hyperspectral imaging reconstruction for deepfake detection, opening new possibilities for detecting increasingly sophisticated manipulations.

cs.CV↗

arXiv2024

Automated Bleeding Detection and Classification in Wireless Capsule Endoscopy with YOLOv8-X

Gastrointestinal (GI) bleeding, a critical indicator of digestive system disorders, re quires efficient and accurate detection methods. This paper presents our solution to the Auto-WCEBleedGen Version V1 Challenge, where we achieved the consolation position. We developed a unified YOLOv8-X model for both detection and classification of bleeding regions in Wireless Capsule Endoscopy (WCE) images. Our approach achieved 96.10% classification accuracy and 76.8% mean Average Precision (mAP) at 0.5 IoU on the val idation dataset. Through careful dataset curation and annotation, we assembled and trained on 6,345 diverse images to ensure robust model performance. Our implementa tion code and trained models are publicly available at https://github.com/pavan98765/Auto-WCEBleedGen.

cs.CV↗