-
Understanding the Challenges in Iterative Generative Optimization with LLMs
Authors:
Allen Nie,
Xavier Daull,
Zhiyi Kuang,
Abhinav Akkiraju,
Anish Chaudhuri,
Max Piasevoli,
Ryan Rong,
YuCheng Yuan,
Prerit Choudhary,
Shannon Xiao,
Rasool Fakoor,
Adith Swaminathan,
Ching-An Cheng
Abstract:
Generative optimization uses large language models (LLMs) to iteratively improve artifacts (such as code, workflows or prompts) using execution feedback. It is a promising approach to building self-improving agents, yet in practice remains brittle: despite active research, only 9% of surveyed agents used any automated optimization. We argue that this brittleness arises because, to set up a learnin…
▽ More
Generative optimization uses large language models (LLMs) to iteratively improve artifacts (such as code, workflows or prompts) using execution feedback. It is a promising approach to building self-improving agents, yet in practice remains brittle: despite active research, only 9% of surveyed agents used any automated optimization. We argue that this brittleness arises because, to set up a learning loop, an engineer must make ``hidden'' design choices: What can the optimizer edit and what is the "right" learning evidence to provide at each update? We investigate three factors that affect most applications: the starting artifact, the credit horizon for execution traces, and batching trials and errors into learning evidence. Through case studies in MLAgentBench, Atari, and BigBench Extra Hard, we find that these design decisions can determine whether generative optimization succeeds, yet they are rarely made explicit in prior work. Different starting artifacts determine which solutions are reachable in MLAgentBench, truncated traces can still improve Atari agents, and larger minibatches do not monotonically improve generalization on BBEH. We conclude that the lack of a simple, universal way to set up learning loops across domains is a major hurdle for productionization and adoption. We provide practical guidance for making these choices.
△ Less
Submitted 25 March, 2026;
originally announced March 2026.
-
Footprint-Guided Exemplar-Free Continual Histopathology Report Generation
Authors:
Pratibha Kumari,
Daniel Reisenbüchler,
Afshin Bozorgpour,
yousef Sadegheih,
Priyankar Choudhary,
Dorit Merhof
Abstract:
Rapid progress in vision-language modeling has enabled pathology report generation from gigapixel whole-slide images, but most approaches assume static training with simultaneous access to all data. In clinical deployment, however, new organs, institutions, and reporting conventions emerge over time, and sequential fine-tuning can cause catastrophic forgetting. We introduce an exemplar-free contin…
▽ More
Rapid progress in vision-language modeling has enabled pathology report generation from gigapixel whole-slide images, but most approaches assume static training with simultaneous access to all data. In clinical deployment, however, new organs, institutions, and reporting conventions emerge over time, and sequential fine-tuning can cause catastrophic forgetting. We introduce an exemplar-free continual learning framework for WSI-to-report generation that avoids storing raw slides or patch exemplars. The core idea is a compact domain footprint built in a frozen patch-embedding space: a small codebook of representative morphology tokens together with slide-level co-occurrence summaries and lightweight patch-count priors. These footprints support generative replay by synthesizing pseudo-WSI representations that reflect domain-specific morphological mixtures, while a teacher snapshot provides pseudo-reports to supervise the updated model without retaining past data. To address shifting reporting conventions, we distill domain-specific linguistic characteristics into a compact style descriptor and use it to steer generation. At inference, the model identifies the most compatible descriptor directly from the slide signal, enabling domain-agnostic setup without requiring explicit domain identifiers. Evaluated across multiple public continual learning benchmarks, our approach outperforms exemplar-free and limited-buffer rehearsal baselines, highlighting footprint-based generative replay as a practical solution for deployment in evolving clinical settings.
△ Less
Submitted 27 February, 2026;
originally announced February 2026.
-
Structured Insight from Unstructured Data: Large Language Models for SDOH-Driven Diabetes Risk Prediction
Authors:
Sasha Ronaghi,
Prerit Choudhary,
David H Rehkopf,
Bryant Lin
Abstract:
Social determinants of health (SDOH) play a critical role in Type 2 Diabetes (T2D) management but are often absent from electronic health records and risk prediction models. Most individual-level SDOH data is collected through structured screening tools, which lack the flexibility to capture the complexity of patient experiences and unique needs of a clinic's population. This study explores the us…
▽ More
Social determinants of health (SDOH) play a critical role in Type 2 Diabetes (T2D) management but are often absent from electronic health records and risk prediction models. Most individual-level SDOH data is collected through structured screening tools, which lack the flexibility to capture the complexity of patient experiences and unique needs of a clinic's population. This study explores the use of large language models (LLMs) to extract structured SDOH information from unstructured patient life stories and evaluate the predictive value of both the extracted features and the narratives themselves for assessing diabetes control. We collected unstructured interviews from 65 T2D patients aged 65 and older, focused on their lived experiences, social context, and diabetes management. These narratives were analyzed using LLMs with retrieval-augmented generation to produce concise, actionable qualitative summaries for clinical interpretation and structured quantitative SDOH ratings for risk prediction modeling. The structured SDOH ratings were used independently and in combination with traditional laboratory biomarkers as inputs to linear and tree-based machine learning models (Ridge, Lasso, Random Forest, and XGBoost) to demonstrate how unstructured narrative data can be applied in conventional risk prediction workflows. Finally, we evaluated several LLMs on their ability to predict a patient's level of diabetes control (low, medium, high) directly from interview text with A1C values redacted. LLMs achieved 60% accuracy in predicting diabetes control levels from interview text. This work demonstrates how LLMs can translate unstructured SDOH-related data into structured insights, offering a scalable approach to augment clinical risk models and decision-making.
△ Less
Submitted 19 January, 2026;
originally announced January 2026.
-
Graph-Based Bayesian Optimization for Quantum Circuit Architecture Search with Uncertainty Calibrated Surrogates
Authors:
Prashant Kumar Choudhary,
Nouhaila Innan,
Muhammad Shafique,
Rajeev Singh
Abstract:
Quantum circuit design is a key bottleneck for practical quantum machine learning on complex, real-world data. We present an automated framework that discovers and refines variational quantum circuits (VQCs) using graph-based Bayesian optimization with a graph neural network (GNN) surrogate. Circuits are represented as graphs and mutated and selected via an expected improvement acquisition functio…
▽ More
Quantum circuit design is a key bottleneck for practical quantum machine learning on complex, real-world data. We present an automated framework that discovers and refines variational quantum circuits (VQCs) using graph-based Bayesian optimization with a graph neural network (GNN) surrogate. Circuits are represented as graphs and mutated and selected via an expected improvement acquisition function informed by surrogate uncertainty with Monte Carlo dropout. Candidate circuits are evaluated with a hybrid quantum-classical variational classifier on the next generation firewall telemetry and network internet of things (NF-ToN-IoT-V2) cybersecurity dataset, after feature selection and scaling for quantum embedding. We benchmark our pipeline against an MLP-based surrogate, random search, and greedy GNN selection. The GNN-guided optimizer consistently finds circuits with lower complexity and competitive or superior classification accuracy compared to all baselines. Robustness is assessed via a noise study across standard quantum noise channels, including amplitude damping, phase damping, thermal relaxation, depolarizing, and readout bit flip noise. The implementation is fully reproducible, with time benchmarking and export of best found circuits, providing a scalable and interpretable route to automated quantum circuit discovery.
△ Less
Submitted 10 December, 2025;
originally announced December 2025.
-
The Service Rate Region of Hamming Codes
Authors:
Priyanka Choudhary,
Maheshanand Bhaintwal
Abstract:
The service rate region of a coded distributed storage system is the set of all achievable data access requests under the capacity constraints. This paper investigates the service rate regions of systematic Hamming codes using hypergraph theory and derives bounds for the maximal achievable service rate of individual data objects. We establish upper bounds on the sum of service rates of data symbol…
▽ More
The service rate region of a coded distributed storage system is the set of all achievable data access requests under the capacity constraints. This paper investigates the service rate regions of systematic Hamming codes using hypergraph theory and derives bounds for the maximal achievable service rate of individual data objects. We establish upper bounds on the sum of service rates of data symbols indexed by a subset of systematic nodes in a systematic binary Hamming code, and explore the achievability of these bounds. Additionally, for non-systematic binary Hamming codes, we conclude that the aggregate service rate is limited by the number of columns of odd weight in the associated generator matrix.
△ Less
Submitted 26 September, 2025;
originally announced September 2025.
-
BountyBench: Dollar Impact of AI Agent Attackers and Defenders on Real-World Cybersecurity Systems
Authors:
Andy K. Zhang,
Joey Ji,
Celeste Menders,
Riya Dulepet,
Thomas Qin,
Ron Y. Wang,
Junrong Wu,
Kyleen Liao,
Jiliang Li,
Jinghan Hu,
Sara Hong,
Nardos Demilew,
Shivatmica Murgai,
Jason Tran,
Nishka Kacheria,
Ethan Ho,
Denis Liu,
Lauren McLane,
Olivia Bruvik,
Dai-Rong Han,
Seungwoo Kim,
Akhil Vyas,
Cuiyuanxiu Chen,
Ryan Li,
Weiran Xu
, et al. (9 additional authors not shown)
Abstract:
AI agents have the potential to significantly alter the cybersecurity landscape. Here, we introduce the first framework to capture offensive and defensive cyber-capabilities in evolving real-world systems. Instantiating this framework with BountyBench, we set up 25 systems with complex, real-world codebases. To capture the vulnerability lifecycle, we define three task types: Detect (detecting a ne…
▽ More
AI agents have the potential to significantly alter the cybersecurity landscape. Here, we introduce the first framework to capture offensive and defensive cyber-capabilities in evolving real-world systems. Instantiating this framework with BountyBench, we set up 25 systems with complex, real-world codebases. To capture the vulnerability lifecycle, we define three task types: Detect (detecting a new vulnerability), Exploit (exploiting a given vulnerability), and Patch (patching a given vulnerability). For Detect, we construct a new success indicator, which is general across vulnerability types and provides localized evaluation. We manually set up the environment for each system, including installing packages, setting up server(s), and hydrating database(s). We add 40 bug bounties, which are vulnerabilities with monetary awards from \$10 to \$30,485, covering 9 of the OWASP Top 10 Risks. To modulate task difficulty, we devise a new strategy based on information to guide detection, interpolating from identifying a zero day to exploiting a given vulnerability. We evaluate 10 agents: Claude Code, OpenAI Codex CLI with o3-high and o4-mini, and custom agents with o3-high, GPT-4.1, Gemini 2.5 Pro Preview, Claude 3.7 Sonnet Thinking, Qwen3 235B A22B, Llama 4 Maverick, and DeepSeek-R1. Given up to three attempts, the top-performing agents are Codex CLI: o3-high (12.5% on Detect, mapping to \$3,720; 90% on Patch, mapping to \$14,152), Custom Agent: Claude 3.7 Sonnet Thinking (67.5% on Exploit), and Codex CLI: o4-mini (90% on Patch, mapping to \$14,422). Codex CLI: o3-high, Codex CLI: o4-mini, and Claude Code are more capable at defense, achieving higher Patch scores of 90%, 90%, and 87.5%, compared to Exploit scores of 47.5%, 32.5%, and 57.5% respectively; while the custom agents are relatively balanced between offense and defense, achieving Exploit scores of 17.5-67.5% and Patch scores of 25-60%.
△ Less
Submitted 2 December, 2025; v1 submitted 21 May, 2025;
originally announced May 2025.
-
Leveraging Large Language Models for Explainable Activity Recognition in Smart Homes: A Critical Evaluation
Authors:
Michele Fiori,
Gabriele Civitarese,
Priyankar Choudhary,
Claudio Bettini
Abstract:
Explainable Artificial Intelligence (XAI) aims to uncover the inner reasoning of machine learning models. In IoT systems, XAI improves the transparency of models processing sensor data from multiple heterogeneous devices, ensuring end-users understand and trust their outputs. Among the many applications, XAI has also been applied to sensor-based Activities of Daily Living (ADLs) recognition in sma…
▽ More
Explainable Artificial Intelligence (XAI) aims to uncover the inner reasoning of machine learning models. In IoT systems, XAI improves the transparency of models processing sensor data from multiple heterogeneous devices, ensuring end-users understand and trust their outputs. Among the many applications, XAI has also been applied to sensor-based Activities of Daily Living (ADLs) recognition in smart homes. Existing approaches highlight which sensor events are most important for each predicted activity, using simple rules to convert these events into natural language explanations for non-expert users. However, these methods produce rigid explanations lacking natural language flexibility and are not scalable. With the recent rise of Large Language Models (LLMs), it is worth exploring whether they can enhance explanation generation, considering their proven knowledge of human activities. This paper investigates potential approaches to combine XAI and LLMs for sensor-based ADL recognition. We evaluate if LLMs can be used: a) as explainable zero-shot ADL recognition models, avoiding costly labeled data collection, and b) to automate the generation of explanations for existing data-driven XAI approaches when training data is available and the goal is higher recognition rates. Our critical evaluation provides insights into the benefits and challenges of using LLMs for explainable ADL recognition.
△ Less
Submitted 20 August, 2025; v1 submitted 20 March, 2025;
originally announced March 2025.
-
HQNN-FSP: A Hybrid Classical-Quantum Neural Network for Regression-Based Financial Stock Market Prediction
Authors:
Prashant Kumar Choudhary,
Nouhaila Innan,
Muhammad Shafique,
Rajeev Singh
Abstract:
Financial time-series forecasting remains a challenging task due to complex temporal dependencies and market fluctuations. This study explores the potential of hybrid quantum-classical approaches to assist in financial trend prediction by leveraging quantum resources for improved feature representation and learning. A custom Quantum Neural Network (QNN) regressor is introduced, designed with a nov…
▽ More
Financial time-series forecasting remains a challenging task due to complex temporal dependencies and market fluctuations. This study explores the potential of hybrid quantum-classical approaches to assist in financial trend prediction by leveraging quantum resources for improved feature representation and learning. A custom Quantum Neural Network (QNN) regressor is introduced, designed with a novel ansatz tailored for financial applications. Two hybrid optimization strategies are proposed: (1) a sequential approach where classical recurrent models (RNN/LSTM) extract temporal dependencies before quantum processing, and (2) a joint learning framework that optimizes classical and quantum parameters simultaneously. Systematic evaluation using TimeSeriesSplit, k-fold cross-validation, and predictive error analysis highlights the ability of these hybrid models to integrate quantum computing into financial forecasting workflows. The findings demonstrate how quantum-assisted learning can contribute to financial modeling, offering insights into the practical role of quantum resources in time-series analysis.
△ Less
Submitted 19 March, 2025;
originally announced March 2025.
-
Online Authentication Habits of Indian Users
Authors:
Pratyush Choudhary,
Subhrajit Das,
Mukul Paras Potta,
Prasuj Das,
Abhishek Bichhawat
Abstract:
Passwords have been long used as the primary authentication method for web services. Weak passwords used by the users have prompted the use of password management tools and two-factor authentication to ensure better account security. While prior studies have studied their adoption individually, none of these studies focuses particularly on the Indian setting, which is culturally and economically d…
▽ More
Passwords have been long used as the primary authentication method for web services. Weak passwords used by the users have prompted the use of password management tools and two-factor authentication to ensure better account security. While prior studies have studied their adoption individually, none of these studies focuses particularly on the Indian setting, which is culturally and economically different from the countries in which these studies have been done in the past. To this end, we conducted a survey with 90 participants residing in India to better understand the mindset of people on using password managers and two-factor authentication (2FA).
Our findings suggest that a majority of the participants have used 2FA and password managers in some form, although they are sometimes unaware of their formal names. While many participants used some form of 2FA across all their accounts, browser-integrated and device-default password managers are predominantly utilized for less sensitive platforms such as e-commerce and social media rather than for more critical accounts like banking. The primary motivation for using password managers is the convenience of auto-filling. However, some participants avoid using password managers due to a lack of trust in these tools. Notably, dedicated third-party applications show low adoption for both password manager and 2FA.
Despite acknowledging the importance of secure password practices, many participants still reuse passwords across multiple accounts, prefer shorter passwords, and use commonly predictable password patterns. Overall, the study suggests that Indians are more inclined to choose default settings, underscoring the need for tailored strategies to improve user awareness and strengthen password security practices.
△ Less
Submitted 24 January, 2025;
originally announced January 2025.
-
Ensuring Fair LLM Serving Amid Diverse Applications
Authors:
Redwan Ibne Seraj Khan,
Kunal Jain,
Haiying Shen,
Ankur Mallick,
Anjaly Parayil,
Anoop Kulkarni,
Steve Kofsky,
Pankhuri Choudhary,
Renèe St. Amant,
Rujia Wang,
Yue Cheng,
Ali R. Butt,
Victor Rühle,
Chetan Bansal,
Saravan Rajmohan
Abstract:
In a multi-tenant large language model (LLM) serving platform hosting diverse applications, some users may submit an excessive number of requests, causing the service to become unavailable to other users and creating unfairness. Existing fairness approaches do not account for variations in token lengths across applications and multiple LLM calls, making them unsuitable for such platforms. To addre…
▽ More
In a multi-tenant large language model (LLM) serving platform hosting diverse applications, some users may submit an excessive number of requests, causing the service to become unavailable to other users and creating unfairness. Existing fairness approaches do not account for variations in token lengths across applications and multiple LLM calls, making them unsuitable for such platforms. To address the fairness challenge, this paper analyzes millions of requests from thousands of users on MS CoPilot, a real-world multi-tenant LLM platform hosted by Microsoft. Our analysis confirms the inadequacy of existing methods and guides the development of FairServe, a system that ensures fair LLM access across diverse applications. FairServe proposes application-characteristic aware request throttling coupled with a weighted service counter based scheduling technique to curb abusive behavior and ensure fairness. Our experimental results on real-world traces demonstrate FairServe's superior performance compared to the state-of-the-art method in ensuring fairness. We are actively working on deploying our system in production, expecting to benefit millions of customers world-wide.
△ Less
Submitted 24 November, 2024;
originally announced November 2024.
-
HDL-GPT: High-Quality HDL is All You Need
Authors:
Bhuvnesh Kumar,
Saurav Nanda,
Ganapathy Parthasarathy,
Pawan Patil,
Austin Tsai,
Parivesh Choudhary
Abstract:
This paper presents Hardware Description Language Generative Pre-trained Transformers (HDL-GPT), a novel approach that leverages the vast repository of open-source High Definition Language (HDL) codes to train superior quality large code models. The core premise of this paper is the hypothesis that high-quality HDL is all you need to create models with exceptional performance and broad zero-shot g…
▽ More
This paper presents Hardware Description Language Generative Pre-trained Transformers (HDL-GPT), a novel approach that leverages the vast repository of open-source High Definition Language (HDL) codes to train superior quality large code models. The core premise of this paper is the hypothesis that high-quality HDL is all you need to create models with exceptional performance and broad zero-shot generalization abilities. The paper elucidates the methods employed for the curation and augmentation of large corpora from open-source HDL code, transforming highly variable quality data into high-quality data through careful prompting and context maintenance. We demonstrate that the careful selection, filtering, and augmentation of data across HDLs can yield powerful models that surpass current state-of-the-art models. We also explore the impact of different fine-tuning methods on the quality of results. We describe experimental results across a range of fine-tuned SOTA LLMs, substantiating our claims. We demonstrate improvements of 50% to 200% over SOTA HDL models on current benchmarks in tasks ranging from HDL circuit explanations, code generation, formal and simulation testbench creation, triaging bugs, and fixing them. HDL-GPT opens new avenues for the development of advanced model training techniques for circuit design tasks.
△ Less
Submitted 25 July, 2024;
originally announced July 2024.
-
Large Language Models are Zero-Shot Recognizers for Activities of Daily Living
Authors:
Gabriele Civitarese,
Michele Fiori,
Priyankar Choudhary,
Claudio Bettini
Abstract:
The sensor-based recognition of Activities of Daily Living (ADLs) in smart home environments enables several applications in the areas of energy management, safety, well-being, and healthcare. ADLs recognition is typically based on deep learning methods requiring large datasets to be trained. Recently, several studies proved that Large Language Models (LLMs) effectively capture common-sense knowle…
▽ More
The sensor-based recognition of Activities of Daily Living (ADLs) in smart home environments enables several applications in the areas of energy management, safety, well-being, and healthcare. ADLs recognition is typically based on deep learning methods requiring large datasets to be trained. Recently, several studies proved that Large Language Models (LLMs) effectively capture common-sense knowledge about human activities. However, the effectiveness of LLMs for ADLs recognition in smart home environments still deserves to be investigated. In this work, we propose ADL-LLM, a novel LLM-based ADLs recognition system. ADLLLM transforms raw sensor data into textual representations, that are processed by an LLM to perform zero-shot ADLs recognition. Moreover, in the scenario where a small labeled dataset is available, ADL-LLM can also be empowered with few-shot prompting. We evaluated ADL-LLM on two public datasets, showing its effectiveness in this domain.
△ Less
Submitted 20 March, 2025; v1 submitted 1 July, 2024;
originally announced July 2024.
-
FusionMind -- Improving question and answering with external context fusion
Authors:
Shreyas Verma,
Manoj Parmar,
Palash Choudhary,
Sanchita Porwal
Abstract:
Answering questions using pre-trained language models (LMs) and knowledge graphs (KGs) presents challenges in identifying relevant knowledge and performing joint reasoning.We compared LMs (fine-tuned for the task) with the previously published QAGNN method for the Question-answering (QA) objective and further measured the impact of additional factual context on the QAGNN performance. The QAGNN met…
▽ More
Answering questions using pre-trained language models (LMs) and knowledge graphs (KGs) presents challenges in identifying relevant knowledge and performing joint reasoning.We compared LMs (fine-tuned for the task) with the previously published QAGNN method for the Question-answering (QA) objective and further measured the impact of additional factual context on the QAGNN performance. The QAGNN method employs LMs to encode QA context and estimate KG node importance, and effectively update the question choice entity representations using Graph Neural Networks (GNNs). We further experimented with enhancing the QA context encoding by incorporating relevant knowledge facts for the question stem. The models are trained on the OpenbookQA dataset, which contains ~6000 4-way multiple choice questions and is widely used as a benchmark for QA tasks. Through our experimentation, we found that incorporating knowledge facts context led to a significant improvement in performance. In contrast, the addition of knowledge graphs to language models resulted in only a modest increase. This suggests that the integration of contextual knowledge facts may be more impactful for enhancing question answering performance compared to solely adding knowledge graphs.
△ Less
Submitted 30 December, 2023;
originally announced January 2024.
-
EXPLORE -- Explainable Song Recommendation
Authors:
Abhinav Arun,
Mehul Soni,
Palash Choudhary,
Saksham Arora
Abstract:
This study explores the development of an explainable music recommendation system with enhanced user control. Leveraging a hybrid of collaborative filtering and content-based filtering, we address the challenges of opaque recommendation logic and lack of user influence on results. We present a novel approach combining advanced algorithms and an interactive user interface. Our methodology integrate…
▽ More
This study explores the development of an explainable music recommendation system with enhanced user control. Leveraging a hybrid of collaborative filtering and content-based filtering, we address the challenges of opaque recommendation logic and lack of user influence on results. We present a novel approach combining advanced algorithms and an interactive user interface. Our methodology integrates Spotify data with user preference analytics to tailor music suggestions. Evaluation through RMSE and user studies underscores the efficacy and user satisfaction with our system. The paper concludes with potential directions for future enhancements in group recommendations and dynamic feedback integration.
△ Less
Submitted 30 December, 2023;
originally announced January 2024.
-
EdgeAISim: A Toolkit for Simulation and Modelling of AI Models in Edge Computing Environments
Authors:
Aadharsh Roshan Nandhakumar,
Ayush Baranwal,
Priyanshukumar Choudhary,
Muhammed Golec,
Sukhpal Singh Gill
Abstract:
To meet next-generation IoT application demands, edge computing moves processing power and storage closer to the network edge to minimise latency and bandwidth utilisation. Edge computing is becoming popular as a result of these benefits, but resource management is still challenging. Researchers are utilising AI models to solve the challenge of resource management in edge computing systems. Howeve…
▽ More
To meet next-generation IoT application demands, edge computing moves processing power and storage closer to the network edge to minimise latency and bandwidth utilisation. Edge computing is becoming popular as a result of these benefits, but resource management is still challenging. Researchers are utilising AI models to solve the challenge of resource management in edge computing systems. However, existing simulation tools are only concerned with typical resource management policies, not the adoption and implementation of AI models for resource management, especially. Consequently, researchers continue to face significant challenges, making it hard and time-consuming to use AI models when designing novel resource management policies for edge computing with existing simulation tools. To overcome these issues, we propose a lightweight Python-based toolkit called EdgeAISim for the simulation and modelling of AI models for designing resource management policies in edge computing environments. In EdgeAISim, we extended the basic components of the EdgeSimPy framework and developed new AI-based simulation models for task scheduling, energy management, service migration, network flow scheduling, and mobility support for edge computing environments. In EdgeAISim, we have utilised advanced AI models such as Multi-Armed Bandit with Upper Confidence Bound, Deep Q-Networks, Deep Q-Networks with Graphical Neural Network, and ActorCritic Network to optimize power usage while efficiently managing task migration within the edge computing environment. The performance of these proposed models of EdgeAISim is compared with the baseline, which uses a worst-fit algorithm-based resource management policy in different settings. Experimental results indicate that EdgeAISim exhibits a substantial reduction in power consumption, highlighting the compelling success of power optimization strategies in EdgeAISim.
△ Less
Submitted 9 October, 2023;
originally announced October 2023.
-
Automatic Signboard Recognition in Low Quality Night Images
Authors:
Manas Kagde,
Priyanka Choudhary,
Rishi Joshi,
Somnath Dey
Abstract:
An essential requirement for driver assistance systems and autonomous driving technology is implementing a robust system for detecting and recognizing traffic signs. This system enables the vehicle to autonomously analyze the environment and make appropriate decisions regarding its movement, even when operating at higher frame rates. However, traffic sign images captured in inadequate lighting and…
▽ More
An essential requirement for driver assistance systems and autonomous driving technology is implementing a robust system for detecting and recognizing traffic signs. This system enables the vehicle to autonomously analyze the environment and make appropriate decisions regarding its movement, even when operating at higher frame rates. However, traffic sign images captured in inadequate lighting and adverse weather conditions are poorly visible, blurred, faded, and damaged. Consequently, the recognition of traffic signs in such circumstances becomes inherently difficult. This paper addressed the challenges of recognizing traffic signs from images captured in low light, noise, and blurriness. To achieve this goal, a two-step methodology has been employed. The first step involves enhancing traffic sign images by applying a modified MIRNet model and producing enhanced images. In the second step, the Yolov4 model recognizes the traffic signs in an unconstrained environment. The proposed method has achieved 5.40% increment in mAP@0.5 for low quality images on Yolov4. The overall mAP@0.5 of 96.75% has been achieved on the GTSRB dataset. It has also attained mAP@0.5 of 100% on the GTSDB dataset for the broad categories, comparable with the state-of-the-art work.
△ Less
Submitted 17 August, 2023;
originally announced August 2023.
-
VAKTA-SETU: A Speech-to-Speech Machine Translation Service in Select Indic Languages
Authors:
Shivam Mhaskar,
Vineet Bhat,
Akshay Batheja,
Sourabh Deoghare,
Paramveer Choudhary,
Pushpak Bhattacharyya
Abstract:
In this work, we present our deployment-ready Speech-to-Speech Machine Translation (SSMT) system for English-Hindi, English-Marathi, and Hindi-Marathi language pairs. We develop the SSMT system by cascading Automatic Speech Recognition (ASR), Disfluency Correction (DC), Machine Translation (MT), and Text-to-Speech Synthesis (TTS) models. We discuss the challenges faced during the research and deve…
▽ More
In this work, we present our deployment-ready Speech-to-Speech Machine Translation (SSMT) system for English-Hindi, English-Marathi, and Hindi-Marathi language pairs. We develop the SSMT system by cascading Automatic Speech Recognition (ASR), Disfluency Correction (DC), Machine Translation (MT), and Text-to-Speech Synthesis (TTS) models. We discuss the challenges faced during the research and development stage and the scalable deployment of the SSMT system as a publicly accessible web service. On the MT part of the pipeline too, we create a Text-to-Text Machine Translation (TTMT) service in all six translation directions involving English, Hindi, and Marathi. To mitigate data scarcity, we develop a LaBSE-based corpus filtering tool to select high-quality parallel sentences from a noisy pseudo-parallel corpus for training the TTMT system. All the data used for training the SSMT and TTMT systems and the best models are being made publicly available. Users of our system are (a) Govt. of India in the context of its new education policy (NEP), (b) tourists who criss-cross the multilingual landscape of India, (c) Indian Judiciary where a leading cause of the pendency of cases (to the order of 10 million as on date) is the translation of case papers, (d) farmers who need weather and price information and so on. We also share the feedback received from various stakeholders when our SSMT and TTMT systems were demonstrated in large public events.
△ Less
Submitted 21 May, 2023;
originally announced May 2023.
-
Concept Drift Challenge in Multimedia Anomaly Detection: A Case Study with Facial Datasets
Authors:
Pratibha Kumari,
Priyankar Choudhary,
Pradeep K. Atrey,
Mukesh Saini
Abstract:
Anomaly detection in multimedia datasets is a widely studied area. Yet, the concept drift challenge in data has been ignored or poorly handled by the majority of the anomaly detection frameworks. The state-of-the-art approaches assume that the data distribution at training and deployment time will be the same. However, due to various real-life environmental factors, the data may encounter drift in…
▽ More
Anomaly detection in multimedia datasets is a widely studied area. Yet, the concept drift challenge in data has been ignored or poorly handled by the majority of the anomaly detection frameworks. The state-of-the-art approaches assume that the data distribution at training and deployment time will be the same. However, due to various real-life environmental factors, the data may encounter drift in its distribution or can drift from one class to another in the late future. Thus, a one-time trained model might not perform adequately. In this paper, we systematically investigate the effect of concept drift on various detection models and propose a modified Adaptive Gaussian Mixture Model (AGMM) based framework for anomaly detection in multimedia data. In contrast to the baseline AGMM, the proposed extension of AGMM remembers the past for a longer period in order to handle the drift better. Extensive experimental analysis shows that the proposed model better handles the drift in data as compared with the baseline AGMM. Further, to facilitate research and comparison with the proposed framework, we contribute three multimedia datasets constituting faces as samples. The face samples of individuals correspond to the age difference of more than ten years to incorporate a longer temporal context.
△ Less
Submitted 27 July, 2022;
originally announced July 2022.
-
On Polynomial Kernels for Traveling Salesperson Problem and its Generalizations
Authors:
Václav Blažej,
Pratibha Choudhary,
Dušan Knop,
Šimon Schierreich,
Ondřej Suchý,
Tomáš Valla
Abstract:
For many problems, the important instances from practice possess certain structure that one should reflect in the design of specific algorithms. As data reduction is an important and inextricable part of today's computation, we employ one of the most successful models of such precomputation -- the kernelization. Within this framework, we focus on Traveling Salesperson Problem (TSP) and some of its…
▽ More
For many problems, the important instances from practice possess certain structure that one should reflect in the design of specific algorithms. As data reduction is an important and inextricable part of today's computation, we employ one of the most successful models of such precomputation -- the kernelization. Within this framework, we focus on Traveling Salesperson Problem (TSP) and some of its generalizations.
We provide a kernel for TSP with size polynomial in either the feedback edge set number or the size of a modulator to constant-sized components. For its generalizations, we also consider other structural parameters such as the vertex cover number and the size of a modulator to constant-sized paths. We complement our results from the negative side by showing that the existence of a polynomial-sized kernel with respect to the fractioning number, the combined parameter maximum degree and treewidth, and, in the case of Subset-TSP, modulator to disjoint cycles (i.e., the treewidth two graphs) is unlikely.
△ Less
Submitted 3 July, 2022;
originally announced July 2022.
-
Polynomial Kernels for Tracking Shortest Paths
Authors:
Václav Blažej,
Pratibha Choudhary,
Dušan Knop,
Jan Matyáš Křišťan,
Ondřej Suchý,
Tomáš Valla
Abstract:
Given an undirected graph $G=(V,E)$, vertices $s,t\in V$, and an integer $k$, Tracking Shortest Paths requires deciding whether there exists a set of $k$ vertices $T\subseteq V$ such that for any two distinct shortest paths between $s$ and $t$, say $P_1$ and $P_2$, we have $T\cap V(P_1)\neq T\cap V(P_2)$. In this paper, we give the first polynomial size kernel for the problem. Specifically we show…
▽ More
Given an undirected graph $G=(V,E)$, vertices $s,t\in V$, and an integer $k$, Tracking Shortest Paths requires deciding whether there exists a set of $k$ vertices $T\subseteq V$ such that for any two distinct shortest paths between $s$ and $t$, say $P_1$ and $P_2$, we have $T\cap V(P_1)\neq T\cap V(P_2)$. In this paper, we give the first polynomial size kernel for the problem. Specifically we show the existence of a kernel with $\mathcal{O}(k^2)$ vertices and edges in general graphs and a kernel with $\mathcal{O}(k)$ vertices and edges in planar graphs for the Tracking Paths in DAG problem. This problem admits a polynomial parameter transformation to Tracking Shortest Paths, and this implies a kernel with $\mathcal{O}(k^4)$ vertices and edges for Tracking Shortest Paths in general graphs and a kernel with $\mathcal{O}(k^2)$ vertices and edges in planar graphs. Based on the above we also give a single exponential algorithm for Tracking Shortest Paths in planar graphs.
△ Less
Submitted 24 February, 2022;
originally announced February 2022.
-
Parameterized Complexity of Minimum Membership Dominating Set
Authors:
Akanksha Agrawal,
Pratibha Choudhary,
N. S. Narayanaswamy,
K. K. Nisha,
Vijayaragunathan Ramamoorthi
Abstract:
Given a graph $G=(V,E)$ and an integer $k$, the Minimum Membership Dominating Set (MMDS) problem seeks to find a dominating set $S \subseteq V$ of $G$ such that for each $v \in V$, $|N[v] \cap S|$ is at most $k$. We investigate the parameterized complexity of the problem and obtain the following results about MMDS:
W[1]-hardness of the problem parameterized by the pathwidth (and thus, treewidth…
▽ More
Given a graph $G=(V,E)$ and an integer $k$, the Minimum Membership Dominating Set (MMDS) problem seeks to find a dominating set $S \subseteq V$ of $G$ such that for each $v \in V$, $|N[v] \cap S|$ is at most $k$. We investigate the parameterized complexity of the problem and obtain the following results about MMDS:
W[1]-hardness of the problem parameterized by the pathwidth (and thus, treewidth) of the input graph.
W[1]-hardness parameterized by $k$ on split graphs.
An algorithm running in time $2^{\mathcal{O}(\textbf{vc})} |V|^{\mathcal{O}(1)}$, where $\textbf{vc}$ is the size of a minimum-sized vertex cover of the input graph.
An ETH-based lower bound showing that the algorithm mentioned in the previous item is optimal.
△ Less
Submitted 13 October, 2021;
originally announced October 2021.
-
Constant Factor Approximation for Tracking Paths and Fault Tolerant Feedback Vertex Set
Authors:
Václav Blažej,
Pratibha Choudhary,
Dušan Knop,
Jan Matyáš Křišťan,
Ondřej Suchý,
Tomáš Valla
Abstract:
Consider a vertex-weighted graph $G$ with a source $s$ and a target $t$. Tracking Paths requires finding a minimum weight set of vertices (trackers) such that the sequence of trackers in each path from $s$ to $t$ is unique. In this work, we derive a factor $6$-approximation algorithm for Tracking Paths in weighted graphs and a factor $4$-approximation algorithm if the input is unweighted. This is…
▽ More
Consider a vertex-weighted graph $G$ with a source $s$ and a target $t$. Tracking Paths requires finding a minimum weight set of vertices (trackers) such that the sequence of trackers in each path from $s$ to $t$ is unique. In this work, we derive a factor $6$-approximation algorithm for Tracking Paths in weighted graphs and a factor $4$-approximation algorithm if the input is unweighted. This is the first constant factor approximation for this problem. While doing so, we also study approximation of the closely related $r$-Fault Tolerant Feedback Vertex Set problem. There, for a fixed integer $r$ and a given vertex-weighted graph $G$, the task is to find a minimum weight set of vertices intersecting every cycle of $G$ in at least $r+1$ vertices. We give a factor $\mathcal{O}(r)$ approximation algorithm for $r$-Fault Tolerant Feedback Vertex Set if $r$ is a constant.
△ Less
Submitted 24 February, 2022; v1 submitted 3 August, 2021;
originally announced August 2021.
-
On Kernels for d-Path Vertex Cover
Authors:
Radovan Červený,
Pratibha Choudhary,
Ondřej Suchý
Abstract:
In this paper we study the kernelization of the $d$-Path Vertex Cover ($d$-PVC) problem. Given a graph $G$, the problem requires finding whether there exists a set of at most $k$ vertices whose removal from $G$ results in a graph that does not contain a path (not necessarily induced) with $d$ vertices. It is known that $d$-PVC is NP-complete for $d\geq 2$. Since the problem generalizes to $d$-Hitt…
▽ More
In this paper we study the kernelization of the $d$-Path Vertex Cover ($d$-PVC) problem. Given a graph $G$, the problem requires finding whether there exists a set of at most $k$ vertices whose removal from $G$ results in a graph that does not contain a path (not necessarily induced) with $d$ vertices. It is known that $d$-PVC is NP-complete for $d\geq 2$. Since the problem generalizes to $d$-Hitting Set, it is known to admit a kernel with $\mathcal{O}(dk^d)$ edges. We improve on this by giving better kernels. Specifically, we give kernels with $\mathcal{O}(k^2)$ vertices and edges for the cases when $d=4$ and $d=5$. Further, we give a kernel with $\mathcal{O}(k^4d^{2d+9})$ vertices and edges for general $d$.
△ Less
Submitted 4 July, 2022; v1 submitted 26 July, 2021;
originally announced July 2021.
-
Structural Parameterizations of Tracking Paths Problem
Authors:
Pratibha Choudhary,
Venkatesh Raman
Abstract:
Given a graph $G$ with source and destination vertices $s,t\in V(G)$ respectively, \textsc{Tracking Paths} asks for a minimum set of vertices $T\subseteq V(G)$, such that the sequence of vertices encountered in each simple path from $s$ to $t$ is unique. The problem was proven \textsc{NP}-hard \cite{tr-j} and was found to admit a quadratic kernel when parameterized by the size of the desired solut…
▽ More
Given a graph $G$ with source and destination vertices $s,t\in V(G)$ respectively, \textsc{Tracking Paths} asks for a minimum set of vertices $T\subseteq V(G)$, such that the sequence of vertices encountered in each simple path from $s$ to $t$ is unique. The problem was proven \textsc{NP}-hard \cite{tr-j} and was found to admit a quadratic kernel when parameterized by the size of the desired solution \cite{quadratic}. Following recent trends, for the first time, we study \textsc{Tracking Paths} with respect to structural parameters of the input graph, parameters that measure how far the input graph is, from an easy instance. We prove that \textsc{Tracking Paths} admits fixed-parameter tractable (\textsc{FPT}) algorithms when parameterized by the size of vertex cover, and the size of cluster vertex deletion set for the input graph.
△ Less
Submitted 22 August, 2020;
originally announced August 2020.
-
Polynomial Time Algorithms for Tracking Path Problems
Authors:
Pratibha Choudhary
Abstract:
Given a graph $G$, and terminal vertices $s$ and $t$, the TRACKING PATHS problem asks to compute a minimum number of vertices to be marked as trackers, such that the sequence of trackers encountered in each s-t path is unique. TRACKING PATHS is NP-hard in both directed and undirected graphs in general. In this paper we give a collection of polynomial time algorithms for some restricted versions of…
▽ More
Given a graph $G$, and terminal vertices $s$ and $t$, the TRACKING PATHS problem asks to compute a minimum number of vertices to be marked as trackers, such that the sequence of trackers encountered in each s-t path is unique. TRACKING PATHS is NP-hard in both directed and undirected graphs in general. In this paper we give a collection of polynomial time algorithms for some restricted versions of TRACKING PATHS. We prove that TRACKING PATHS is polynomial time solvable for chordal graphs and tournament graphs. We prove that TRACKING PATHS is NP-hard in graphs with bounded maximum degree $δ\geq 6$, and give a $2(δ+1)$-approximate algorithm for the same. We also analyze the version of tracking s-t paths where paths are tracked using edges instead of vertices, and we give a polynomial time algorithm for the same. Finally, we show how to reconstruct an s-t path, given a sequence of trackers and a tracking set for the graph in consideration.
△ Less
Submitted 18 February, 2020;
originally announced February 2020.
-
Fixed-parameter tractable algorithms for Tracking Shortest Paths
Authors:
Aritra Banik,
Pratibha Choudhary,
Venkatesh Raman,
Saket Saurabh
Abstract:
We consider the parameterized complexity of the problem of tracking shortest s-t paths in graphs, motivated by applications in security and wireless networks. Given an undirected and unweighted graph with a source s and a destination t, Tracking Shortest Paths asks if there exists a k-sized subset of vertices (referred to as tracking set) that intersects each shortest s-t path in a distinct set of…
▽ More
We consider the parameterized complexity of the problem of tracking shortest s-t paths in graphs, motivated by applications in security and wireless networks. Given an undirected and unweighted graph with a source s and a destination t, Tracking Shortest Paths asks if there exists a k-sized subset of vertices (referred to as tracking set) that intersects each shortest s-t path in a distinct set of vertices. We first generalize this problem for set systems, namely Tracking Set System, where given a family of subsets of a universe, we are required to find a subset of elements from the universe that has a unique intersection with each set in the family. Tracking Set System is shown to be fixed-parameter tractable due to its relation with a known problem, Test Cover. By a reduction to the well-studied d-hitting set problem, we give a polynomial (with respect to k) kernel for the case when the set sizes are bounded by d. This also helps solving Tracking Shortest Paths when the input graph diameter is bounded by d. While the results for Tracking Set System help to show that Tracking Shortest Paths is fixed-parameter tractable, we also give an independent algorithm by using some preprocessing rules, resulting in an improved running time.
△ Less
Submitted 18 August, 2020; v1 submitted 24 January, 2020;
originally announced January 2020.
-
Improved Kernels for Tracking Path Problem
Authors:
Pratibha Choudhary,
Venkatesh Raman
Abstract:
Tracking of moving objects is crucial to security systems and networks. Given a graph $G$, terminal vertices $s$ and $t$, and an integer $k$, the \textsc{Tracking Paths} problem asks whether there exists at most $k$ vertices, which if marked as trackers, would ensure that the sequence of trackers encountered in each s-t path is unique. It is known that the problem is NP-hard and admits a kernel (r…
▽ More
Tracking of moving objects is crucial to security systems and networks. Given a graph $G$, terminal vertices $s$ and $t$, and an integer $k$, the \textsc{Tracking Paths} problem asks whether there exists at most $k$ vertices, which if marked as trackers, would ensure that the sequence of trackers encountered in each s-t path is unique. It is known that the problem is NP-hard and admits a kernel (reducible to an equivalent instance) with $\mathcal{O}(k^6)$ vertices and $\mathcal{O}(k^7)$ edges, when parameterized by the size of the output (tracking set) $k$ [5]. An interesting question that remains open is whether the existing kernel can be improved. In this paper we answer this affirmatively: (i) For general graphs, we show the existence of a kernel of size $\mathcal{O}(k^2)$, (ii) For planar graphs, we improve this further by giving a kernel of size $\mathcal{O}(k)$. In addition, we also show that finding a tracking set of size at most $n-k$ for a graph on $n$ vertices is hard for the parameterized complexity class W[1], when parameterized by $k$.
△ Less
Submitted 21 August, 2020; v1 submitted 9 January, 2020;
originally announced January 2020.
-
Data-Driven Investigative Journalism For Connectas Dataset
Authors:
Aniket Jain,
Bhavya Sharma,
Paridhi Choudhary,
Rohan Sangave,
William Yang
Abstract:
The following paper explores the possibility of using Machine Learning algorithms to detect the cases of corruption and malpractice by governments. The dataset used by the authors contains information about several government contracts in Colombia from year 2007 to 2012. The authors begin with exploring and cleaning the data, followed by which they perform feature engineering before finally implem…
▽ More
The following paper explores the possibility of using Machine Learning algorithms to detect the cases of corruption and malpractice by governments. The dataset used by the authors contains information about several government contracts in Colombia from year 2007 to 2012. The authors begin with exploring and cleaning the data, followed by which they perform feature engineering before finally implementing Machine Learning models to detect anomalies in the given dataset.
△ Less
Submitted 23 April, 2018;
originally announced April 2018.
-
HTTPI Based Web Service Security over SOAP
Authors:
Pankaj Choudhary,
Rajendra Aaseri,
Nirmal Roberts
Abstract:
Now a days, a new family of web applications open applications, are emerging (e.g., Social Networking, News and Blogging). Generally, these open applications are non-confidential. The security needs of these applications are only client/server authentication and data integrity. For securing these open applications, effectively and efficiently, HTTPI, a new transport protocol is proposed, which ens…
▽ More
Now a days, a new family of web applications open applications, are emerging (e.g., Social Networking, News and Blogging). Generally, these open applications are non-confidential. The security needs of these applications are only client/server authentication and data integrity. For securing these open applications, effectively and efficiently, HTTPI, a new transport protocol is proposed, which ensures the entire security requirements of open applications. Benefit of using the HTTPI is that it is economical in use, well-suited for cache proxies, like HTTP is, and provides security against many Internet attacks (Server Impersonation and Message Modification) like HTTPS does. In terms of performance HTTPI is very close to the HTTP, but much better than HTTPS. A Web service is a method of communication between two ends over the Internet. These web services are developed over XML and HTTP. Today, most of the open applications use web services for most of their operations. For securing these web services, security design based on HTTPI is proposed. Our work involves securing the web services over SOAP, based on the HTTPI. This secure web service might be applicable for open applications, where authentication and integrity is needed, but no confidentiality required. In our paper, we introduce a web service security model based on HTTPI protocol over SOAP and develop a preliminary implementation of this model. We also analyze the performance of our approach through an experiment and show that our proposed approach provides higher throughput, lower average response time and lower response size than HTTPS based web service security approach.
△ Less
Submitted 7 June, 2013;
originally announced June 2013.