arXiv subjects

Priyanshu Gupta

Publications and source records attributed to Priyanshu Gupta.

16 recordsLinked to original sources

arXiv2026

Back to the Future: A workbook time machine for spread sheet creation benchmarks

We introduce the workbook time machine, a pipeline that automatically creates benchmarks evaluating the ability of language models to create derived objects in spreadsheets (formulas, charts, pivot tables, and conditional formatting). Applied to public workbook corpora, it produces wtmcorpus--a collection of (input workbook, output workbook, query) triples spanning four artifact types and varying complexity. From this corpus we curate wtmbench, a 150-task evaluation benchmark with queries at three levels of specificity. We evaluate existing spreadsheet manipulation agents and baselines on wtmbench across artifact types, step complexity, and instruction granularity. Our evaluations show that query specificity, agent orchestration, and interface API used to control spreadsheets play a big role in LLM performance on Excel tasks.

cs.AI↗

arXiv2026

Reason Wide, Not Deep: Amortizing the Reasoning Premium into Distilled Skills

Reasoning modes of language models outperform their non-reasoning counterparts on multi-step agentic tasks, but pay a 3-6x premium in output tokens on every episode -- much of it spent re-deriving procedures that are shared across episodes of the same domain. We show this recurring cost can be amortized: a coding agent analyses a small corpus of existing trajectories from a training split and compiles a compact natural-language skill that is injected into the non-reasoning model's system prompt. Across four agentic benchmarks (ALFWorld, tau$^2$-bench telecom and retail, and SpreadsheetBench-Verified), skills recover 55%-100%+ of the reasoning gap for GPT-5.4-mini on held-out tasks -- exceeding the reasoning mode outright on two of four -- while emitting 2.7-6x fewer output tokens and zero reasoning tokens. Notably, reasoning traces are not a prerequisite: skills distilled from non-reasoning trajectories alone remain competitive with skills distilled from paired reasoning/non-reasoning corpora, with domain-dependent differences between the two sources. We interpret these results through a search lens: test-time reasoning is deep search inside a single episode, re-paid at every deployment, while corpus distillation is wide search across episodes, paid once. The two recover overlapping procedural knowledge, and width over cheap trajectories is often the better buy -- with the residual gap on some domains (telecom, SpreadsheetBench) delineating where genuinely per-instance deep search remains necessary.

cs.AI↗

arXiv2026

Virtual receiver functions via conditional diffusion transformers for robust crustal imaging

Receiver functions (RFs) are widely used to image crustal and upper-mantle structure, and their variation with backazimuth and epicentral distance contains key information about layering and azimuthal anisotropy. In practice, however, RFs are contaminated by nuisance effects from unknown earthquake source signatures and seismic noise, which obstruct reliable crustal imaging. Sparse RF coverage across backazimuths and epicentral distances also leads to biased anisotropy estimates. We address these challenges using conditional diffusion models, conditioned on backazimuth, epicentral distance, and station coordinates, to produce high-quality virtual radial and transverse RFs. RFs from earthquakes with similar backazimuths and epicentral distances share consistent crustal responses but differ in nuisance effects, allowing the model to suppress the latter. Our framework generates virtual RFs within gaps in backazimuth and epicentral distance coverage, enhancing the interpretation of crustal anisotropy and layering. On synthetic RFs with realistic non-Gaussian noise, virtual RFs correlate more strongly with the true RFs than traditional linear or phase-weighted stacking. Applied to the Cascadia Subduction Zone, virtual radial RFs sharply image scattered S-waves from the dipping slab, with enhanced phase clarity and backazimuthal coverage relative to previous work. In southern California, anisotropy parameters inferred from virtual RFs are spatially coherent and consistent with regional fault geometry. Our approach leverages all available RFs, regardless of quality, to increase spatial coverage and support robust, automated RF analysis.

physics.geo-ph↗

arXiv2025

Improving Language Agents through BREW: Bootstrapping expeRientially-learned Environmental knoWledge

Large Language Model (LLM)-based agents are increasingly capable of complex, multi-step tasks such as GUI automation, tool use, and data manipulation, yet they cannot learn from experience: each new session rediscovers solutions from scratch. We introduce BREW (Bootstrapping expeRientially-learned Environmental knoWledge), a framework that distills an agent's past interaction trajectories into a structured, retrievable knowledge base (KB) of natural-language recipes, concept-level procedural documents that capture what to do, when it applies, and what to watch out for. Drawing on the principle of library learning from program synthesis, BREW decomposes agent memory into modular, concept-localized documents and formalizes KB construction as a state-space search problem. To navigate this space, we introduce Expand-and-Gather Monte Carlo Tree Search (EG-MCTS), a reward-guided algorithm that jointly optimizes recipe accuracy and retrievability across parallel, per-concept search trees. We further adapt hindsight relabeling to convert near-miss trajectories into positive demonstrations, surfacing latent agent competencies as reusable knowledge. On three domain-grounded benchmarks, OSWorld, tau^2-Bench, and SpreadSheetBench, BREW achieves 10-20% gains in task success and 10-15% fewer execution steps over base agents, while consistently outperforming existing memory-augmented baselines that can degrade below memoryless performance. The resulting KB is inspectable, modular, and extensible, providing a transparent and controllable substrate for agent optimization.

cs.AI↗

arXiv2025

Distributions In, Distributions Out: The Case for Soft-Label Training

Supervised classifiers output a distribution over classes but are typically trained against a single label obtained by collapsing multiple annotators into a majority vote. On tasks where annotator disagreement reflects genuine ambiguity -- natural language inference, politeness, visually ambiguous categorization -- this collapse discards information and forces models to express uniform confidence on inputs where humans systematically disagree. We compare soft-label training, which uses the full annotation distribution as the target, against hard-label training across three datasets spanning vision and NLP (ChaosNLI, POPQUORN, CIFAR-10H). Soft-label training matches or exceeds hard-label accuracy on every dataset, reduces KL divergence to the annotator distribution by 32% on average (p < 10^-4), and produces predictions whose per-sample entropy correlates 61% more strongly with annotator entropy -- models trained on distributions are uncertain precisely where humans are. We argue these benefits follow from a basic observation: when annotators legitimately disagree, the annotation distribution is the correct learning target, not a noisy estimate of it.

cs.LG↗

arXiv2025

STACKFEED: Structured Textual Actor-Critic Knowledge Base Editing with FeedBack

Large Language Models (LLMs) often generate incorrect or outdated information, especially in low-resource settings or when dealing with private data. To address this, Retrieval-Augmented Generation (RAG) uses external knowledge bases (KBs), but these can also suffer from inaccuracies. We introduce STACKFEED, a novel Structured Textual Actor-Critic Knowledge base editing with FEEDback approach that iteratively refines the KB based on expert feedback using a multi-actor, centralized critic reinforcement learning framework. STACKFEED defines a ReACT actor agent on each document to perform structured edits based on document specific targeted instructions. Experimental results showcase that STACKFEED significantly improves KB quality and performance of the RAG system. We evaluate STACKFEED on low-resource programming problems, modified python packaged and factual question-answering tasks.

cs.AI↗

arXiv2025

ConDABench: Interactive Evaluation of Language Models for Data Analysis

Real-world data analysis tasks often come with under-specified goals and unclean data. User interaction is necessary to understand and disambiguate a user's intent, and hence, essential to solving these complex tasks. Existing benchmarks for evaluating LLMs on data analysis tasks do not capture these complexities or provide first-class support for interactivity. We introduce ConDABench, a framework for generating conversational data analysis (ConDA) benchmarks and evaluating external tools on the generated benchmarks. \bench consists of (a) a multi-agent workflow for generating realistic benchmarks from articles describing insights gained from public datasets, (b) 1,420 ConDA problems generated using this workflow, and (c) an evaluation harness that, for the first time, makes it possible to systematically evaluate conversational data analysis tools on the generated ConDA problems. Evaluation of state-of-the-art LLMs on the benchmarks reveals that while the new generation of models are better at solving more instances, they are not necessarily better at solving tasks that require sustained, long-form engagement. ConDABench is an avenue for model builders to measure progress towards truly collaborative models that can complete complex interactive tasks.

cs.CL↗

arXiv2025

PU-Lie: Lightweight Deception Detection in Imbalanced Diplomatic Dialogues via Positive-Unlabeled Learning

Detecting deception in strategic dialogues is a complex and high-stakes task due to the subtlety of language and extreme class imbalance between deceptive and truthful communications. In this work, we revisit deception detection in the Diplomacy dataset, where less than 5% of messages are labeled deceptive. We introduce a lightweight yet effective model combining frozen BERT embeddings, interpretable linguistic and game-specific features, and a Positive-Unlabeled (PU) learning objective. Unlike traditional binary classifiers, PU-Lie is tailored for situations where only a small portion of deceptive messages are labeled, and the majority are unlabeled. Our model achieves a new best macro F1 of 0.60 while reducing trainable parameters by over 650x. Through comprehensive evaluations and ablation studies across seven models, we demonstrate the value of PU learning, linguistic interpretability, and speaker-aware representations. Notably, we emphasize that in this problem setting, accurately detecting deception is more critical than identifying truthful messages. This priority guides our choice of PU learning, which explicitly models the rare but vital deceptive class.

cs.CL↗

arXiv2024

MetaReflection: Learning Instructions for Language Agents using Past Reflections

The popularity of Large Language Models (LLMs) have unleashed a new age ofLanguage Agents for solving a diverse range of tasks. While contemporary frontier LLMs are capable enough to power reasonably good Language agents, the closed-API model makes it hard to improve in cases they perform sub-optimally. To address this, recent works have explored ways to improve their performance using techniques like self-reflection and prompt optimization. Unfortunately, techniques like self-reflection can be used only in an online setup, while contemporary prompt optimization techniques are designed and tested to work on simple tasks. To this end, we introduce MetaReflection, a novel offline reinforcement learning technique that enhances the performance of Language Agents by augmenting a semantic memory based on experiential learnings from past trials. We demonstrate the efficacy of MetaReflection by evaluating across multiple domains, including complex logical reasoning, biomedical semantic similarity, open world question answering, and vulnerability threat detection, in Infrastructure-as-Code, spanning different agent designs. MetaReflection boosts Language agents' performance by 4% to 16.82% over the raw GPT-4 baseline and performs on par with existing state-of-the-art prompt optimization techniques while requiring fewer LLM calls.

cs.CL↗

arXiv2024

LOGIC-LM++: Multi-Step Refinement for Symbolic Formulations

In this paper we examine the limitations of Large Language Models (LLMs) for complex reasoning tasks. Although recent works have started to employ formal languages as an intermediate representation for reasoning tasks, they often face challenges in accurately generating and refining these formal specifications to ensure correctness. To address these issues, this paper proposes Logic-LM++, an improvement on Logic-LM . It uses the ability of LLMs to do pairwise comparisons, allowing the evaluation of the refinements suggested by the LLM. The paper demonstrates that Logic-LM++ outperforms Logic-LM and other contemporary techniques across natural language reasoning tasks on three datasets, FOLIO, ProofWriter and AR-LSAT, with an average improvement of 18.5% on standard prompting, 12.3% on chain of thought prompting and 5% on Logic-LM.

cs.LO↗

arXiv2023

TST$^\mathrm{R}$: Target Similarity Tuning Meets the Real World

Target similarity tuning (TST) is a method of selecting relevant examples in natural language (NL) to code generation through large language models (LLMs) to improve performance. Its goal is to adapt a sentence embedding model to have the similarity between two NL inputs match the similarity between their associated code outputs. In this paper, we propose different methods to apply and improve TST in the real world. First, we replace the sentence transformer with embeddings from a larger model, which reduces sensitivity to the language distribution and thus provides more flexibility in synthetic generation of examples, and we train a tiny model that transforms these embeddings to a space where embedding similarity matches code similarity, which allows the model to remain a black box and only requires a few matrix multiplications at inference time. Second, we show how to efficiently select a smaller number of training examples to train the TST model. Third, we introduce a ranking-based evaluation for TST that does not require end-to-end code generation experiments, which can be expensive to perform.

cs.AI↗

arXiv2023

Augmented Embeddings for Custom Retrievals

Information retrieval involves selecting artifacts from a corpus that are most relevant to a given search query. The flavor of retrieval typically used in classical applications can be termed as homogeneous and relaxed, where queries and corpus elements are both natural language (NL) utterances (homogeneous) and the goal is to pick most relevant elements from the corpus in the Top-K, where K is large, such as 10, 25, 50 or even 100 (relaxed). Recently, retrieval is being used extensively in preparing prompts for large language models (LLMs) to enable LLMs to perform targeted tasks. These new applications of retrieval are often heterogeneous and strict -- the queries and the corpus contain different kinds of entities, such as NL and code, and there is a need for improving retrieval at Top-K for small values of K, such as K=1 or 3 or 5. Current dense retrieval techniques based on pretrained embeddings provide a general-purpose and powerful approach for retrieval, but they are oblivious to task-specific notions of similarity of heterogeneous artifacts. We introduce Adapted Dense Retrieval, a mechanism to transform embeddings to enable improved task-specific, heterogeneous and strict retrieval. Adapted Dense Retrieval works by learning a low-rank residual adaptation of the pretrained black-box embedding. We empirically validate our approach by showing improvements over the state-of-the-art general-purpose embeddings-based baseline.

cs.IR↗

arXiv2023

GrACE: Generation using Associated Code Edits

Developers expend a significant amount of time in editing code for a variety of reasons such as bug fixing or adding new features. Designing effective methods to predict code edits has been an active yet challenging area of research due to the diversity of code edits and the difficulty of capturing the developer intent. In this work, we address these challenges by endowing pre-trained large language models (LLMs) of code with the knowledge of prior, relevant edits. The generative capability of the LLMs helps address the diversity in code changes and conditioning code generation on prior edits helps capture the latent developer intent. We evaluate two well-known LLMs, Codex and CodeT5, in zero-shot and fine-tuning settings respectively. In our experiments with two datasets, the knowledge of prior edits boosts the performance of the LLMs significantly and enables them to generate 29% and 54% more correctly edited code in top-1 suggestions relative to the current state-of-the-art symbolic and neural approaches, respectively.

cs.SE↗

arXiv2022

Overwatch: Learning Patterns in Code Edit Sequences

Integrated Development Environments (IDEs) provide tool support to automate many source code editing tasks. Traditionally, IDEs use only the spatial context, i.e., the location where the developer is editing, to generate candidate edit recommendations. However, spatial context alone is often not sufficient to confidently predict the developer's next edit, and thus IDEs generate many suggestions at a location. Therefore, IDEs generally do not actively offer suggestions and instead, the developer is usually required to click on a specific icon or menu and then select from a large list of potential suggestions. As a consequence, developers often miss the opportunity to use the tool support because they are not aware it exists or forget to use it. To better understand common patterns in developer behavior and produce better edit recommendations, we can additionally use the temporal context, i.e., the edits that a developer was recently performing. To enable edit recommendations based on temporal context, we present Overwatch, a novel technique for learning edit sequence patterns from traces of developers' edits performed in an IDE. Our experiments show that Overwatch has 78% precision and that Overwatch not only completed edits when developers missed the opportunity to use the IDE tool support but also predicted new edits that have no tool support in the IDE.

cs.PL↗

arXiv2021

A Probabilistic Framework for Knowledge Graph Data Augmentation

We present NNMFAug, a probabilistic framework to perform data augmentation for the task of knowledge graph completion to counter the problem of data scarcity, which can enhance the learning process of neural link predictors. Our method can generate potentially diverse triples with the advantage of being efficient and scalable as well as agnostic to the choice of the link prediction model and dataset used. Experiments and analysis done on popular models and benchmarks show that NNMFAug can bring notable improvements over the baselines.

cs.LG↗

arXiv2020

Unbiased Loss Functions for Extreme Classification With Missing Labels

The goal in extreme multi-label classification (XMC) is to tag an instance with a small subset of relevant labels from an extremely large set of possible labels. In addition to the computational burden arising from large number of training instances, features and labels, problems in XMC are faced with two statistical challenges, (i) large number of 'tail-labels' -- those which occur very infrequently, and (ii) missing labels as it is virtually impossible to manually assign every relevant label to an instance. In this work, we derive an unbiased estimator for general formulation of loss functions which decompose over labels, and then infer the forms for commonly used loss functions such as hinge- and squared-hinge-loss and binary cross-entropy loss. We show that the derived unbiased estimators, in the form of appropriate weighting factors, can be easily incorporated in state-of-the-art algorithms for extreme classification, thereby scaling to datasets with hundreds of thousand labels. However, empirically, we find a slightly altered version that gives more relative weight to tail labels to perform even better. We suspect is due to the label imbalance in the dataset, which is not explicitly addressed by our theoretically derived estimator. Minimizing the proposed loss functions leads to significant improvement over existing methods (up to 20% in some cases) on benchmark datasets in XMC.

stat.ML↗