arXiv subjects

Sainyam Galhotra

Publications and source records attributed to Sainyam Galhotra.

At least 19 recordsLinked to original sources

arXiv2026

How Far Are We From True Auto-Research?

Recent auto-research systems can produce complete papers, but feasibility is not the same as quality, and the field still lacks a systematic study of how good agent-generated papers actually are. We introduce ResearchArena, a minimal scaffold that lets off-the-shelf agents (Claude Code using Opus 4.6, Codex using GPT-5.4, and Kimi Code using K2.5) carry out the full research loop themselves (ideation, experimentation, paper writing, self-refinement) under only lightweight guidance. Across 13 computer science seeds and 3 trials per agent-domain pair, ResearchArena yields 117 agent-generated papers, each evaluated under three complementary lenses: a manuscript-only reviewer (SAR), an artifact-aware peer review (PR) in which agents inspect the workspace alongside the manuscript, and an human conducted meta-review. Under SAR alone the picture is optimistic: Claude Code obtains the highest score, outperforms Analemma's FARS, and matches the weighted-average human ICLR 2025 submission, suggesting that minimally scaffolded agents can produce papers that look competitive on manuscript-only review. Manual inspection, however, reveals this picture is overstated: SAR scores are poorly aligned with its actual acceptance decisions and reward plausible framing without verifying experimental substance. Under artifact-aware PR scores drop sharply, and manual auditing identifies experimental rigor as the major bottleneck, decomposing into three failure modes (fabricated results, underpowered experiments, and plan/execution mismatch) that are highly agent-dependent: Codex 5%/8% paper-vs-artifact mismatch / fabricated references versus Kimi Code 77%/72%, a $\sim$15$\times$ spread that tracks distinct research personas the agents develop. None of the 117 agent-generated papers reaches the acceptance bar of a top-tier venue. This suggests that we are still gapped from the true auto-research.

cs.AI↗

arXiv2026

HARP: Efficient Data Selection for Finetuning Large Language Models

Finetuning data selection requires balancing two competing goals: selecting examples that improve the downstream objective, and doing so without repeatedly finetuning models. Train-free selectors are scalable but rely on proxies such as embedding similarity or clustering, which may not match the target objective. Train-based selectors better reflect downstream utility through gradient signals, subset evaluation, or Shapley attribution, but require many costly train-evaluate iterations. We propose Hierarchical Active Region Pruning (HARP), an efficient train-based selector that preserves downstream alignment while reducing selection cost. HARP organizes the training pool into a node-leaf hierarchy, evaluates only representative leaves, and infers unmeasured utilities with empirical Bayes posteriors. It then selects data using two complementary envelopes: HARP-C, which conservatively controls redundancy, and HARP-E, which additively rewards complementary regions. We theoretically show that, under local smoothness and bounded estimation error, HARP controls selection error while reducing train-evaluate cost. Across 3 base models, 3 finetuning datasets, and 4 reasoning benchmarks (36 settings), HARP variants achieve the highest accuracy among ranked methods in 23 settings and obtain 36-setting mean accuracies of 64.2 and 63.5, respectively, compared with 59.3 for the strongest baseline, NICE. HARP-C trains on roughly 9x fewer examples than the budgeted baselines and 75x fewer than full finetuning, combining higher mean accuracy with a smaller mean training-set size than every ranked baseline.

cs.LG↗

arXiv2026

MetaJoin: Adaptive Join Selections for Heterogeneous Tables

Joining heterogeneous tables requires different forms of correspondence, including exact agreement, syntactic transformation, semantic similarity, and entity-level matching. The appropriate join method is often unknown in advance, while running every candidate method incurs unnecessary cost. We introduce MetaJoin, a label-free framework for adaptive join selection. MetaJoin uses input evidence to choose an initial specialist, then uses specialist-specific output evidence to determine whether that result should be retained, compared with an alternative, escalated, or repaired. This allows join selection to adapt after execution rather than committing to a one-shot routing decision. We also introduce MNN Join, a transformation specialist that searches for string transformations by maximizing mutual-nearest-neighbor coverage between two unaligned columns and uses the same reciprocal structure to validate its output. We evaluate MetaJoin on 670 datasets from six benchmarks. It achieves 0.817 macro row-F1, slightly exceeding a benchmark-level oracle that is given the best standalone specialist for each benchmark (0.812), and outperforms label-free selection policies built from the same specialist portfolio by more than 0.13. Across the 609 datasets in the four transformation-dominant benchmarks, MNN Join outperforms every evaluated transformation baseline. Finally, MetaJoin's accuracy varies by only 0.015 as the controller backbone ranges from 3B to 14B parameters, showing that output validation and repair substantially reduce dependence on the initial routing decision.

cs.DB↗

arXiv2026

Target-Constrained Pooling for Trajectory Extrapolation

Trajectory extrapolation from a partial target trajectory is underdetermined, but completed related trajectories provide information about the unresolved continuation. We propose target-constrained pooling (TCP), which separates the finite-horizon endpoint from continuation shape and estimates that endpoint from two complementary source signals. Uniform pooling captures final response levels, while matched-increment pooling captures the remaining change after aligning source and target states. TCP selects representative sources, projects both endpoint estimates onto a target-consistent region, converts them into corrections of the free forecast, and controls their influence through strength calibration and degradation-aware weighting. Across 25,245 targets from five domains and both forecasting directions, TCP reduces mean RMSE by 35.8% to 42.1% for all ten target-only forecasters and improves all 100 forecaster, dataset, and direction comparisons. The two endpoint signals consistently outperform either alone, while representative selection remains within 0.64% to 1.02% of the full donor pool using 12.7x fewer donor inputs and 138.8x fewer matching distance computations.

cs.LG↗

arXiv2026

CAFÉ: Causal Black-Box Testing of Machine Unlearning

Machine learning models are increasingly deployed as software components that must evolve as requirements change. When specific training records or features must no longer influence a deployed model, machine unlearning aims to remove that influence without retraining from scratch. Because unlearning is often approximate, its effectiveness must be tested. Such tests must often treat the model as a black box, without access to its parameters, training history, or unlearning procedure. Features pose a further challenge: even after a feature is removed from a model's inputs, its influence can persist through downstream features. Many existing checks examine only the feature's direct use and can therefore certify a model that still depends on it. We frame unlearning testing as specification-based testing and present CAFÉ, which, using only a deployed model's predictions, intervenes on the feature, propagates the change to its downstream features, and checks whether the predictions still respond. CAFÉ measures a target's residual influence through both its direct and indirect causal paths, and its fine-grained diagnostics show which channels and subgroups still carry it. On two causal-network benchmarks with four unlearning methods, CAFÉ ranks residual influence with 0.92--0.93 pairwise accuracy, against at most 0.71 for existing checks, which fail in both directions: they certify models whose influence persists through downstream features and flag correctly unlearned ones. On real census data, CAFÉ likewise exposes influence that survives retraining yet goes unnoticed by direct-input checks.

cs.SE↗

arXiv2026

SALUS: Automated Auditing of NL-to-SQL Benchmarks through Weak Supervision of Multi-Agent Output

Natural language to SQL (NL-to-SQL) benchmarks are foundational to progress in data analysis research, yet recent work has shown that widely-used benchmarks contain significant annotation errors. These errors silently corrupt evaluation metrics, penalize correct model output, and distort the field's understanding of state-of-the-art performance. We present SALUS, a system that automatically detects annotation errors in NL-to-SQL benchmarks. SALUS frames benchmark auditing as a weakly supervised error detection: SQL generated by multiple LLM agents drive a suite of complementary weak-labeling functions. By passing this noisy vote matrix through a generative label model, we extract high-confidence training samples without requiring human ground truth. These samples train a decision plane that maps gold SQL query features to per-agent trustworthiness, allowing SALUS to intelligently fuse reliability estimates with raw verdicts for rigorous benchmark error detection. We evaluate on BIRD-Clean-xs, a benchmark of 298 BIRD development tasks with manually verified correctness labels. SALUS achieves F1 = 0.9194, significantly outperforming the state-of-the-art baselines. Applying SALUS to the full development sets, we estimate annotation error rates of approximately 37% on BIRD and 27% on Spider.

cs.DB↗

arXiv2026

Shifting Mechanisms: How Positional Encoding Choice Shapes In-Context Retrieval

Language models increasingly use architectures that vary attention span and positional encoding across layers, such as applying RoPE with sliding-window attention and NoPE with global attention (SWA NoPE). However, how these choices shape in-context retrieval remains unclear. To study this question, we take a mechanistic view, tracing how positional encoding (PE) choice shapes the internal mechanisms models use for in-context retrieval. Across 22 open-weight models spanning eight families, we find that standard RoPE models rely primarily on positional retrieval, while PE hybrids shift toward semantic retrieval. We further show on a controlled pre-training ablation that confining positional encoding to local layers produces this semantic shift, degrading representations of positional information. Finally, we show that the reported long-context gains of PE hybrids mask a retrieval trade-off: SWA NoPE improves over RoPE on multiple-target retrieval and QA, but degrades when distinguishing competing keys. We show that these behavioral differences better track the mechanism shift from positional toward semantic mechanisms than a uniform improvement in long-context retrieval.

cs.CL↗

arXiv2026

Trace-Level Analysis of Information Contamination in Multi-Agent Systems

Reasoning over heterogeneous artifacts (PDFs, spreadsheets, slide decks, etc.) increasingly occurs within structured agent workflows that iteratively extract, transform, and reference external information. In these workflows, uncertainty is not merely an input-quality issue: it can redirect decomposition and routing decisions, reshape intermediate state, and produce qualitatively different execution trajectories. We study this phenomenon by treating uncertainty as a controlled variable: we inject structured perturbations into artifact-derived representations, execute fixed workflows under comprehensive logging, and quantify contamination via trace divergence in plans, tool invocations, and intermediate state. Across 614 paired runs on 32 GAIA tasks with three different language models, we find a decoupling: workflows may diverge substantially yet recover correct answers, or remain structurally similar while producing incorrect outputs. We characterize three manifestation types: silent semantic corruption, behavioral detours with recovery, and combined structural disruption and their control-flow signatures (rerouting, extended execution, early termination). We measure operational costs and characterize why commonly used verification guardrails fail to intercept contamination. We contribute (i) a formal taxonomy of contamination manifestations in structured workflows, (ii) a trace-based measurement framework for detecting and localizing contamination across agent interactions, and (iii) empirical evidence with implications for targeted verification, defensive design, and cost control.

cs.AI↗

arXiv2026

Towards Reliable Testing of Machine Unlearning

Machine learning components are now central to AI-infused software systems, from recommendations and code assistants to clinical decision support. As regulations and governance frameworks increasingly require deleting sensitive data from deployed models, machine unlearning is emerging as a practical alternative to full retraining. However, unlearning introduces a software quality-assurance challenge: under realistic deployment constraints and imperfect oracles, how can we test that a model no longer relies on targeted information? This paper frames unlearning testing as a first-class software engineering problem. We argue that practical unlearning tests must provide (i) thorough coverage over proxy and mediated influence pathways, (ii) debuggable diagnostics that localize where leakage persists, (iii) cost-effective regression-style execution under query budgets, and (iv) black-box applicability for API-deployed models. We outline a causal, pathway-centric perspective, causal fuzzing, that generates budgeted interventions to estimate residual direct and indirect effects and produce actionable "leakage reports". Proof-of-concept results illustrate that standard attribution checks can miss residual influence due to proxy pathways, cancellation effects, and subgroup masking, motivating causal testing as a promising direction for unlearning testing.

cs.SE↗

arXiv2026

Debugging Concept Bottleneck Models through Removal and Retraining

Concept Bottleneck Models (CBMs) use a set of human-interpretable concepts to predict the final task label, enabling domain experts to not only validate the CBM's predictions, but also intervene on incorrect concepts at test time. However, these interventions fail to address systemic misalignment between the CBM and the expert's reasoning, such as when the model learns shortcuts from biased data. To address this, we present a general interpretable debugging framework for CBMs that follows a two-step process of Removal and Retraining. In the Removal step, experts use concept explanations to identify and remove any undesired concepts. In the Retraining step, we introduce CBDebug, a novel method that leverages the interpretability of CBMs as a bridge for converting concept-level user feedback into sample-level auxiliary labels. These labels are then used to apply supervised bias mitigation and targeted augmentation, reducing the model's reliance on undesired concepts. We evaluate our framework with both real and automated expert feedback, and find that CBDebug significantly outperforms prior retraining methods across multiple CBM architectures (PIP-Net, Post-hoc CBM) and benchmarks with known spurious correlations.

cs.CV↗

arXiv2026

Pruning Minimal Reasoning Graphs for Efficient Retrieval-Augmented Generation

Retrieval-augmented generation (RAG) is now standard for knowledge-intensive LLM tasks, but most systems still treat every query as fresh, repeatedly re-retrieving long passages and re-reasoning from scratch, inflating tokens, latency, and cost. We present AutoPrunedRetriever, a graph-style RAG system that persists the minimal reasoning subgraph built for earlier questions and incrementally extends it for later ones. AutoPrunedRetriever stores entities and relations in a compact, ID-indexed codebook and represents questions, facts, and answers as edge sequences, enabling retrieval and prompting over symbolic structure instead of raw text. To keep the graph compact, we apply a two-layer consolidation policy (fast ANN/KNN alias detection plus selective $k$-means once a memory threshold is reached) and prune low-value structure, while prompts retain only overlap representatives and genuinely new evidence. We instantiate two front ends: AutoPrunedRetriever-REBEL, which uses REBEL as a triplet parser, and AutoPrunedRetriever-llm, which swaps in an LLM extractor. On GraphRAG-Benchmark (Medical and Novel), both variants achieve state-of-the-art complex reasoning accuracy, improving over HippoRAG2 by roughly 9--11 points, and remain competitive on contextual summarize and generation. On our harder STEM and TV benchmarks, AutoPrunedRetriever again ranks first, while using up to two orders of magnitude fewer tokens than graph-heavy baselines, making it a practical substrate for long-running sessions, evolving corpora, and multi-agent pipelines.

cs.DB↗

arXiv2026

Metric $k$-clustering using only Weak Comparison Oracles

Clustering is a fundamental primitive in unsupervised learning. However, classical algorithms for $k$-clustering (such as $k$-median and $k$-means) assume access to exact pairwise distances -- an unrealistic requirement in many modern applications. We study clustering in the \emph{Rank-model (R-model)}, where access to distances is entirely replaced by a \emph{quadruplet oracle} that provides only relative distance comparisons. In practice, such an oracle can represent learned models or human feedback, and is expected to be noisy and entail an access cost. Given a metric space with $n$ input items, we design randomized algorithms that, using only a noisy quadruplet oracle, compute a set of $O(k \cdot \mathsf{polylog}(n))$ centers along with a mapping from the input items to the centers such that the clustering cost of the mapping is at most constant times the optimum $k$-clustering cost. Our method achieves a query complexity of $O(n\cdot k \cdot \mathsf{polylog}(n))$ for arbitrary metric spaces and improves to $O((n+k^2) \cdot \mathsf{polylog}(n))$ when the underlying metric has bounded doubling dimension. When the metric has bounded doubling dimension we can further improve the approximation from constant to $1+\varepsilon$, for any arbitrarily small constant $\varepsilon\in(0,1)$, while preserving the same asymptotic query complexity. Our framework demonstrates how noisy, low-cost oracles, such as those derived from large language models, can be systematically integrated into scalable clustering algorithms.

cs.LG↗

arXiv2026

Octopus: A Lightweight Entity-Aware System for Multi-Table Data Discovery and Cell-Level Retrieval

Tabular data constitute a dominant form of information in modern data lakes and repositories, yet discovering the relevant tables to answer user questions remains challenging. Existing data discovery systems assume that each question can be answered by a single table and often rely on resource-intensive offline preprocessing, such as model training or large-scale content indexing. In practice, however, many questions require information spread across multiple tables -- either independently or through joins -- and users often seek specific cell values rather than entire tables. In this paper, we present Octopus, a lightweight, entity-aware, and training-free system for multi-table data discovery and cell-level value retrieval. Instead of embedding entire questions, Octopus identifies fine-grained entities (column mentions and value mentions) from natural-language queries using an LLM parser. It then matches these entities to table headers through a compact embedding index and scans table contents directly for value occurrences, eliminating the need for heavy content indexing or costly offline stages. The resulting fine-grained alignment not only improves table retrieval accuracy but also facilitates efficient downstream NL2SQL execution by reducing token usage and redundant LLM calls. To evaluate Octopus, we introduce a new benchmark covering both table- and cell-level discovery under multi-table settings, including five datasets for independent discovery and two for join-based discovery. Experimental results show that Octopus consistently outperforms existing systems while achieving substantially lower computational and token costs. Code is available at https://github.com/wenzhilics/octopus.

cs.DB↗

arXiv2025

VERIRAG: A Post-Retrieval Auditing of Scientific Study Summaries

Can democratized information gatekeepers and community note writers effectively decide what scientific information to amplify? Lacking domain expertise, such gatekeepers rely on automated reasoning agents that use RAG to ground evidence to cited sources. But such standard RAG systems validate summaries via semantic grounding and suffer from "methodological blindness," treating all cited evidence as equally valid regardless of rigor. To address this, we introduce VERIRAG, a post-retrieval auditing framework that shifts the task from classification to methodological vulnerability detection. Using private Small Language Models (SLMs), VERIRAG audits source papers against the Veritable taxonomy of statistical rigor. We contribute: (1) a benchmark of 1,730 summaries with realistic, non-obvious perturbations modeled after retracted papers; (2) the auditable Veritable taxonomy; and (3) an operational system that improves Macro F1 by at least 19 points over baselines using GPT-based SLMs, a result that replicates across MISTRAL and Gemma architectures. Given the complexity of detecting non-obvious flaws, we view VERIRAG as a "vulnerability-detection copilot," providing structured audit trails for human editors. In our experiments, individual human testers found over 80% of the generated audit trails useful for decision-making. We plan to release the dataset and code to support responsible science advocacy.

cs.IR↗

arXiv2025

Benchmarking Robust Aggregation in Decentralized Gradient Marketplaces

The rise of distributed and privacy-preserving machine learning has sparked interest in decentralized gradient marketplaces, where participants trade intermediate artifacts like gradients. However, existing Federated Learning (FL) benchmarks overlook critical economic and systemic factors unique to such marketplaces-cost-effectiveness, fairness to sellers, and market stability-especially when a buyer relies on a private baseline dataset for evaluation. We introduce a comprehensive benchmark framework to holistically evaluate robust gradient aggregation methods within these buyer-baseline-reliant marketplaces. Our contributions include: (1) a simulation environment modeling marketplace dynamics with a variable buyer baseline and diverse seller distributions; (2) an evaluation methodology augmenting standard FL metrics with marketplace-centric dimensions such as Economic Efficiency, Fairness, and Selection Dynamics; (3) an in-depth empirical analysis of the existing Distributed Gradient Marketplace framework, MartFL, including the integration and comparative evaluation of adapted FLTrust and SkyMask as alternative aggregation strategies within it. This benchmark spans diverse datasets, local attacks, and Sybil attacks targeting the marketplace selection process; and (4) actionable insights into the trade-offs between model performance, robustness, cost, fairness, and stability. This benchmark equips the community with essential tools and empirical evidence to evaluate and design more robust, equitable, and economically viable decentralized gradient marketplaces.

cs.LG↗

arXiv2025

Towards General-Purpose Data Discovery: A Programming Languages Approach

Efficient and effective data discovery is critical for many modern applications in machine learning and data science. One major bottleneck to the development of a general-purpose data discovery tool is the absence of an expressive formal language, and corresponding implementation, for characterizing and solving generic discovery queries. To this end, we present TQL, a domain-specific language for data discovery well-designed to leverage and exploit the results of programming languages research in both its syntax and semantics. In this paper, we fully and formally characterize the core language through an algebraic model, Imperative Relational Algebra with Types (ImpRAT), and implement a modular proof-of-concept system prototype.

cs.DB↗

arXiv2025

TQL: Towards Type-Driven Data Discovery

Existing query languages for data discovery exhibit system-driven designs that emphasize database features and functionality over user needs. We propose a re-prioritization of the client through an introduction of a language-driven approach to data discovery systems that can leverage powerful results from programming languages research. In this paper, we describe TQL, a flexible and practical query language which incorporates a type-like system to encompass downstream transformation-context in its discovery queries. The syntax and semantics of TQL (including the underlying evaluation model), are formally defined, and a sketch of its implementation is also provided. Additionally, we provide comparisons to existing languages for data retrieval and data discovery to examine the advantages of TQL's expanded expressive power in real-life settings.

cs.DB↗

arXiv2025

VeriMinder: Mitigating Analytical Vulnerabilities in NL2SQL

Application systems using natural language interfaces to databases (NLIDBs) have democratized data analysis. This positive development has also brought forth an urgent challenge to help users who might use these systems without a background in statistical analysis to formulate bias-free analytical questions. Although significant research has focused on text-to-SQL generation accuracy, addressing cognitive biases in analytical questions remains underexplored. We present VeriMinder, https://veriminder.ai, an interactive system for detecting and mitigating such analytical vulnerabilities. Our approach introduces three key innovations: (1) a contextual semantic mapping framework for biases relevant to specific analysis contexts (2) an analytical framework that operationalizes the Hard-to-Vary principle and guides users in systematic data analysis (3) an optimized LLM-powered system that generates high-quality, task-specific prompts using a structured process involving multiple candidates, critic feedback, and self-reflection. User testing confirms the merits of our approach. In direct user experience evaluation, 82.5% participants reported positively impacting the quality of the analysis. In comparative evaluation, VeriMinder scored significantly higher than alternative approaches, at least 20% better when considered for metrics of the analysis's concreteness, comprehensiveness, and accuracy. Our system, implemented as a web application, is set to help users avoid "wrong question" vulnerability during data analysis. VeriMinder code base with prompts, https://reproducibility.link/veriminder, is available as an MIT-licensed open-source software to facilitate further research and adoption within the community.

cs.CL↗