arXiv subjects

Ayush Garg

Publications and source records attributed to Ayush Garg.

14 recordsLinked to original sources

arXiv2026

SEEK: Semantic Evidence Extraction via Adaptive ChunKing for Multilingual Fact-Checking

Multilingual fact verification requires evidence that is both relevant and sufficiently complete for reliable factuality prediction. However, existing systems often rely on search snippets, sentence-level evidence, or locally segmented passages, which can miss decisive context and produce fragmented evidence. To overcome these limitations, we propose SEEK, a Semantic Evidence Extraction with an adaptive chunKing framework that constructs coherent evidence chunks from full fact-checking articles by identifying semantic topic transitions and preserving local verification context. The constructed chunks are encoded using a multilingual encoder and then multilingual LLMs are finetuned using LoRA adapter for veracity prediction. Experiments on X-FACT and RU22Fact show that SEEK improves macro-f1 by up to 10% over semantic chunking, 19% over sentence chunking, and 20% over search-snippet baselines. Evidence completeness and significance analyses further show that SEEK preserves richer verification context and enables more reliable multilingual fact-checking.

cs.CL↗

arXiv2026

RuleForge: Automated Generation and Validation for Web Vulnerability Detection at Scale

Security teams face a challenge: the volume of newly disclosed Common Vulnerabilities and Exposures (CVEs) far exceeds the capacity to manually develop detection mechanisms. In 2025, the National Vulnerability Database published over 48,000 new vulnerabilities, motivating the need for automation. We present RuleForge, an AWS internal system that automatically generates detection rules--JSON-based patterns that identify malicious HTTP requests exploiting specific vulnerabilities--from structured Nuclei templates describing CVE details. Nuclei templates provide standardized, YAML-based vulnerability descriptions that serve as the structured input for our rule generation process. This paper focuses on RuleForge's architecture and operational deployment for CVE-related threat detection, with particular emphasis on our novel LLM-as-a-judge (Large Language Model as judge) confidence validation system and systematic feedback integration mechanism. This validation approach evaluates candidate rules across two dimensions--sensitivity (avoiding false negatives) and specificity (avoiding false positives)--achieving AUROC of 0.75 and reducing false positives by 67% compared to synthetic-test-only validation in production. Our 5x5 generation strategy (five parallel candidates with up to five refinement attempts each) combined with continuous feedback loops enables systematic quality improvement. We also present extensions enabling rule generation from unstructured data sources and demonstrate a proof-of-concept agentic workflow for multi-event-type detection. Our lessons learned highlight critical considerations for applying LLMs to cybersecurity tasks, including overconfidence mitigation and the importance of domain expertise in both prompt design and quality review of generated rules through human-in-the-loop validation.

cs.CR↗

arXiv2025

Mind the Links: Cross-Layer Attention for Link Prediction in Multiplex Networks

Multiplex graphs capture diverse relations among shared nodes. Most predictors either collapse layers or treat them independently. This loses crucial inter-layer dependencies and struggles with scalability. To overcome this, we frame multiplex link prediction as multi-view edge classification. For each node pair, we construct a sequence of per-layer edge views and apply cross-layer self-attention to fuse evidence for the target layer. We present two models as instances of this framework: Trans-SLE, a lightweight transformer over static embeddings, and Trans-GAT, which combines layer-specific GAT encoders with transformer fusion. To ensure scalability and fairness, we introduce a Union--Set candidate pool and two leakage-free protocols: cross-layer and inductive subgraph generalization. Experiments on six public multiplex datasets show consistent macro-F_1 gains over strong baselines (MELL, HOPLP-MUL, RMNE). Our approach is simple, scalable, and compatible with both precomputed embeddings and GNN encoders.

cs.LG↗

arXiv2025

Entailed Opinion Matters: Improving the Fact-Checking Performance of Language Models by Relying on their Entailment Ability

Automated fact-checking has been a challenging task for the research community. Prior work has explored various strategies, such as end-to-end training, retrieval-augmented generation, and prompt engineering, to build robust fact-checking systems. However, their accuracy has not been high enough for real-world deployment. We, on the other hand, propose a new learning paradigm, where evidence classification and entailed justifications made by generative language models (GLMs) are used to train encoder-only language models (ELMs). We conducted a rigorous set of experiments, comparing our approach with recent works along with various prompting and fine-tuning strategies. Additionally, we performed ablation studies, error analysis, quality analysis of model explanations, and a domain generalisation study to provide a comprehensive understanding of our approach.

cs.CL↗

arXiv2024

Graph Transformers without Positional Encodings

Recently, Transformers for graph representation learning have become increasingly popular, achieving state-of-the-art performance on a wide-variety of graph datasets, either alone or in combination with message-passing graph neural networks (MP-GNNs). Infusing graph inductive-biases in the innately structure-agnostic transformer architecture in the form of structural or positional encodings (PEs) is key to achieving these impressive results. However, designing such encodings is tricky and disparate attempts have been made to engineer such encodings including Laplacian eigenvectors, relative random-walk probabilities (RRWP), spatial encodings, centrality encodings, edge encodings etc. In this work, we argue that such encodings may not be required at all, provided the attention mechanism itself incorporates information about the graph structure. We introduce Eigenformer, a Graph Transformer employing a novel spectrum-aware attention mechanism cognizant of the Laplacian spectrum of the graph, and empirically show that it achieves performance competetive with SOTA Graph Transformers on a number of standard GNN benchmarks. Additionally, we theoretically prove that Eigenformer can express various graph structural connectivity matrices, which is particularly essential when learning over smaller graphs.

cs.LG↗

arXiv2023

Towards Earlier Detection of Oral Diseases On Smartphones Using Oral and Dental RGB Images

Oral diseases such as periodontal (gum) diseases and dental caries (cavities) affect billions of people across the world today. However, previous state-of-the-art models have relied on X-ray images to detect oral diseases, making them inaccessible to remote monitoring, developing countries, and telemedicine. To combat this overuse of X-ray imagery, we propose a lightweight machine learning model capable of detecting calculus (also known as hardened plaque or tartar) in RGB images while running efficiently on low-end devices. The model, a modified MobileNetV3-Small neural network transfer learned from ImageNet, achieved an accuracy of 72.73% (which is comparable to state-of-the-art solutions) while still being able to run on mobile devices due to its reduced memory requirements and processing times. A ResNet34-based model was also constructed and achieved an accuracy of 81.82%. Both of these models were tested on a mobile app, demonstrating their potential to limit the number of serious oral disease cases as their predictions can help patients schedule appointments earlier without the need to go to the clinic.

cs.CV↗

arXiv2022

Convex geometries representable with colors, by ellipses on the plane, and impossible by circles

A convex geometry is a closure system satisfying the anti-exchange property. This paper, following the work of K. Adaricheva and M. Bolat (2016) and the Polymath REU 2020 team, continues to investigate representations of convex geometries on a 5-element base set. It introduces several properties: the opposite property, nested triangle property, area Q property, and separation property, of convex geometries of circles on a plane, preventing this representation for numerous convex geometries on a 5-element base set. It also demonstrates that all 672 convex geometries on a 5-element base set have a representation by ellipses, as given in the appendix for those without a known representation by circles, and introduces a method of expanding representation with circles by defining unary predicates, shown as colors.

math.CO↗

arXiv2022

Isolating Functions at the Hardware Limit with Virtines

An important class of applications, including programs that leverage third-party libraries, programs that use user-defined functions in databases, and serverless applications, benefit from isolating the execution of untrusted code at the granularity of individual functions or function invocations. However, existing isolation mechanisms were not designed for this use case; rather, they have been adapted to it. We introduce \textit{virtines}, a new abstraction designed specifically for function granularity isolation, and describe how we build virtines from the ground up by pushing hardware virtualization to its limits. Virtines give developers fine-grained control in deciding which functions should run in isolated environments, and which should not. The virtine abstraction is a general one, and we demonstrate a prototype that adds extensions to the C language. We present a detailed analysis of the overheads of running individual functions in isolated VMs, and guided by those findings, we present Wasp, an embeddable hypervisor that allows programmers to easily use virtines. We describe several representative scenarios that employ individual function isolation, and demonstrate that virtines can be applied in these scenarios with only a few lines of changes to existing codebases and with acceptable slowdowns.

cs.PL↗

arXiv2021

Convex geometries representable by at most 5 circles on the plane

A convex geometry is a closure system satisfying the anti-exchange property. In this work we document all convex geometries on 4- and 5-element base sets with respect to their representation by circles on the plane. All 34 non-isomorphic geometries on a 4-element set can be represented by circles, and of the 672 geometries on a 5-element set, we made representations of 623. Of the 49 remaining geometries on a 5-element set, one was already shown not to be representable due to the Weak Carousel property, as articulated by Adaricheva and Bolat (Discrete Mathematics, 2019). In this paper we show that 7 more of these convex geometries cannot be represented by circles on the plane, due to what we term the Triangle Property.

math.CO↗

arXiv2021

Towards efficient end-to-end speech recognition with biologically-inspired neural networks

Automatic speech recognition (ASR) is a capability which enables a program to process human speech into a written form. Recent developments in artificial intelligence (AI) have led to high-accuracy ASR systems based on deep neural networks, such as the recurrent neural network transducer (RNN-T). However, the core components and the performed operations of these approaches depart from the powerful biological counterpart, i.e., the human brain. On the other hand, the current developments in biologically-inspired ASR models, based on spiking neural networks (SNNs), lag behind in terms of accuracy and focus primarily on small scale applications. In this work, we revisit the incorporation of biologically-plausible models into deep learning and we substantially enhance their capabilities, by taking inspiration from the diverse neural and synaptic dynamics found in the brain. In particular, we introduce neural connectivity concepts emulating the axo-somatic and the axo-axonic synapses. Based on this, we propose novel deep learning units with enriched neuro-synaptic dynamics and integrate them into the RNN-T architecture. We demonstrate for the first time, that a biologically realistic implementation of a large-scale ASR model can yield competitive performance levels compared to the existing deep learning models. Specifically, we show that such an implementation bears several advantages, such as a reduced computational cost and a lower latency, which are critical for speech recognition applications.

eess.AS↗

arXiv2021

MIPE: A Metric Independent Pipeline for Effective Code-Mixed NLG Evaluation

Code-mixing is a phenomenon of mixing words and phrases from two or more languages in a single utterance of speech and text. Due to the high linguistic diversity, code-mixing presents several challenges in evaluating standard natural language generation (NLG) tasks. Various widely popular metrics perform poorly with the code-mixed NLG tasks. To address this challenge, we present a metric independent evaluation pipeline MIPE that significantly improves the correlation between evaluation metrics and human judgments on the generated code-mixed text. As a use case, we demonstrate the performance of MIPE on the machine-generated Hinglish (code-mixing of Hindi and English languages) sentences from the HinGE corpus. We can extend the proposed evaluation strategy to other code-mixed language pairs, NLG tasks, and evaluation metrics with minimal to no effort.

cs.CL↗

arXiv2020

A Machine Learning Application for Raising WASH Awareness in the Times of COVID-19 Pandemic

Background: The COVID-19 pandemic has uncovered the potential of digital misinformation in shaping the health of nations. The deluge of unverified information that spreads faster than the epidemic itself is an unprecedented phenomenon that has put millions of lives in danger. Mitigating this Infodemic requires strong health messaging systems that are engaging, vernacular, scalable, effective and continuously learn the new patterns of misinformation. Objective: We created WashKaro, a multi-pronged intervention for mitigating misinformation through conversational AI, machine translation and natural language processing. WashKaro provides the right information matched against WHO guidelines through AI, and delivers it in the right format in local languages. Methods: We theorize (i) an NLP based AI engine that could continuously incorporate user feedback to improve relevance of information, (ii) bite sized audio in the local language to improve penetrance in a country with skewed gender literacy ratios, and (iii) conversational but interactive AI engagement with users towards an increased health awareness in the community. Results: A total of 5026 people who downloaded the app during the study window, among those 1545 were active users. Our study shows that 3.4 times more females engaged with the App in Hindi as compared to males, the relevance of AI-filtered news content doubled within 45 days of continuous machine learning, and the prudence of integrated AI chatbot Satya increased thus proving the usefulness of an mHealth platform to mitigate health misinformation. Conclusion: We conclude that a multi-pronged machine learning application delivering vernacular bite-sized audios and conversational AI is an effective approach to mitigate health misinformation.

cs.CY↗

arXiv2020

SEAL: Scientific Keyphrase Extraction and Classification

Automatic scientific keyphrase extraction is a challenging problem facilitating several downstream scholarly tasks like search, recommendation, and ranking. In this paper, we introduce SEAL, a scholarly tool for automatic keyphrase extraction and classification. The keyphrase extraction module comprises two-stage neural architecture composed of Bidirectional Long Short-Term Memory cells augmented with Conditional Random Fields. The classification module comprises of a Random Forest classifier. We extensively experiment to showcase the robustness of the system. We evaluate multiple state-of-the-art baselines and show a significant improvement. The current system is hosted at http://lingo.iitgn.ac.in:5000/.

cs.IR↗

arXiv2019

BioGen: Automated Biography Generation

A biography of a person is the detailed description of several life events including his education, work, relationships, and death. Wikipedia, the free web-based encyclopedia, consists of millions of manually curated biographies of eminent politicians, film and sports personalities, etc. However, manual curation efforts, even though efficient, suffers from significant delays. In this work, we propose an automatic biography generation framework BioGen. BioGen generates a short collection of biographical sentences clustered into multiple events of life. Evaluation results show that biographies generated by BioGen are significantly closer to manually written biographies in Wikipedia. A working model of this framework is available at nlpbiogen.herokuapp.com/home/.

cs.DL↗