Selected work

Representative Papers

CosMAP: Contrastive Manifold Approximation and Projection for Dimensionality Reduction of Omics and Genealogical Data

Aug 10, 2026

This work addresses the challenge of simultaneously preserving local neighborhoods, global structure, and population coherence when reducing the dimensionality of high-dimensional, sparse omics and lineage data. To this end, the authors propose a graph-based unsupervised dimensionality reduction method that constructs an initial neighborhood graph using cosine similarity and optimizes an attraction–repulsion objective in the embedding space via temperature-normalized contrastive affinities. A two-stage optimization strategy is introduced: first, an intermediate high-dimensional representation is used to refine the neighborhood graph and initialize the embedding; second, the final low-dimensional representation is fine-tuned. Evaluated on single-cell RNA-seq, handwritten digit, and large-scale lineage datasets, the method consistently outperforms existing approaches, yielding more coherent visualizations, superior neighborhood preservation, and clearer global structural organization.

0 citationsRead paper

Transformer-Guided Swarm Intelligence for Frugal Neural Architecture Search

Jul 13, 2026

This work addresses the high computational cost, cold-start issues, and model bloat commonly associated with traditional neural architecture search (NAS). To overcome these limitations, the authors propose a frugal membrane-inspired NAS framework that integrates an autoregressive Transformer controller for global macro-search with an artificial bee colony (ABC) algorithm for local micro-optimization. A dynamic entropy mechanism is introduced to enhance topological exploration and mitigate premature convergence, while explicit regularization of network depth controls parameter scale. The method efficiently discovers a lightweight model on CIFAR-10 within three hours, achieving 84.85% accuracy with only 174K parameters, and attains an F1 score of 0.71 on a credit card fraud detection task using approximately 4.6K parameters, significantly improving the efficiency and practicality of NAS for edge devices.

0 citationsRead paper

Towards Robust Artificial Intelligence: Self-supervised Learning Approach for Out-of-Distribution Detection

Oct 14, 2025ARES

To address the insufficient robustness of AI systems against out-of-distribution (OOD) samples in safety-critical applications (e.g., autonomous driving, healthcare), this paper proposes a fully unsupervised, self-supervised OOD detection method that requires no labeled data. Our approach innovatively integrates self-supervised representation learning with graph-theoretic structural analysis to model semantic relationships among samples in an unsupervised manner, enabling high-accuracy discrimination of OOD instances. Specifically, it learns discriminative features solely from unlabeled data and identifies OOD samples via graph connectivity metrics derived from the learned representations. Evaluated on standard benchmarks, the method achieves an AUROC of 0.99—substantially surpassing current state-of-the-art unsupervised approaches—and marks the first time near-supervised-level OOD detection performance has been attained under completely label-free conditions. This work establishes a scalable, low-dependency paradigm for robust AI deployment.

0 citationsRead paper

FaaSGuard: Secure CI/CD for Serverless Applications -- An OpenFaaS Case Study

Sep 04, 2025

To address the lack of end-to-end security protection in CI/CD pipelines for open-source Serverless platforms (e.g., OpenFaaS), this paper proposes the first unified DevSecOps pipeline tailored for open-source Serverless environments. Our approach integrates static code analysis, dynamic behavior monitoring, secret scanning, and resource-constraint policies to deliver lightweight, fault-closed security checks, deeply embedded across all CI/CD stages. Unlike existing ad hoc security practices, our work provides the first systematic, low-overhead defense spanning the entire Serverless development lifecycle—effectively mitigating common threats such as injection attacks and hardcoded secrets. Evaluation on 20 real-world Serverless functions demonstrates 95% precision and 91% recall in vulnerability detection, with negligible runtime overhead.

0 citationsRead paper

A Data-Centric Approach for Safe and Secure Large Language Models against Threatening and Toxic Content

Apr 19, 2025

To address toxicity, bias, and jailbreaking risks inherent in large language model (LLM) outputs, this paper proposes BART-Corrective—a data-driven, model-agnostic post-hoc toxicity correction framework. Unlike approaches requiring architectural modifications, prompt engineering, or fine-tuning, BART-Corrective employs a lightweight, pre-trained BART-based correction model to purify LLM generations in real time. It introduces a “data-centric” post-processing paradigm, integrating multi-dimensional toxicity detection (e.g., Perspective API) and enabling plug-and-play compatibility across diverse LLMs. Extensive evaluation on four major LLMs—GPT-4, PaLM2, Mistral, and Gemma—demonstrates an average 28% reduction in toxicity and up to a 23% decrease in jailbreak success rate. The framework significantly enhances safety, generalizability, and deployment flexibility without compromising generation quality or inference efficiency.

0 citationsRead paper

Recent publications

Latest Papers

CosMAP: Contrastive Manifold Approximation and Projection for Dimensionality Reduction of Omics and Genealogical Data

Aug 10, 2026

This work addresses the challenge of simultaneously preserving local neighborhoods, global structure, and population coherence when reducing the dimensionality of high-dimensional, sparse omics and lineage data. To this end, the authors propose a graph-based unsupervised dimensionality reduction method that constructs an initial neighborhood graph using cosine similarity and optimizes an attraction–repulsion objective in the embedding space via temperature-normalized contrastive affinities. A two-stage optimization strategy is introduced: first, an intermediate high-dimensional representation is used to refine the neighborhood graph and initialize the embedding; second, the final low-dimensional representation is fine-tuned. Evaluated on single-cell RNA-seq, handwritten digit, and large-scale lineage datasets, the method consistently outperforms existing approaches, yielding more coherent visualizations, superior neighborhood preservation, and clearer global structural organization.

0 citationsRead paper

Transformer-Guided Swarm Intelligence for Frugal Neural Architecture Search

Jul 13, 2026

This work addresses the high computational cost, cold-start issues, and model bloat commonly associated with traditional neural architecture search (NAS). To overcome these limitations, the authors propose a frugal membrane-inspired NAS framework that integrates an autoregressive Transformer controller for global macro-search with an artificial bee colony (ABC) algorithm for local micro-optimization. A dynamic entropy mechanism is introduced to enhance topological exploration and mitigate premature convergence, while explicit regularization of network depth controls parameter scale. The method efficiently discovers a lightweight model on CIFAR-10 within three hours, achieving 84.85% accuracy with only 174K parameters, and attains an F1 score of 0.71 on a credit card fraud detection task using approximately 4.6K parameters, significantly improving the efficiency and practicality of NAS for edge devices.

0 citationsRead paper

Towards Robust Artificial Intelligence: Self-supervised Learning Approach for Out-of-Distribution Detection

Oct 14, 2025ARES

To address the insufficient robustness of AI systems against out-of-distribution (OOD) samples in safety-critical applications (e.g., autonomous driving, healthcare), this paper proposes a fully unsupervised, self-supervised OOD detection method that requires no labeled data. Our approach innovatively integrates self-supervised representation learning with graph-theoretic structural analysis to model semantic relationships among samples in an unsupervised manner, enabling high-accuracy discrimination of OOD instances. Specifically, it learns discriminative features solely from unlabeled data and identifies OOD samples via graph connectivity metrics derived from the learned representations. Evaluated on standard benchmarks, the method achieves an AUROC of 0.99—substantially surpassing current state-of-the-art unsupervised approaches—and marks the first time near-supervised-level OOD detection performance has been attained under completely label-free conditions. This work establishes a scalable, low-dependency paradigm for robust AI deployment.

0 citationsRead paper

FaaSGuard: Secure CI/CD for Serverless Applications -- An OpenFaaS Case Study

Sep 04, 2025

To address the lack of end-to-end security protection in CI/CD pipelines for open-source Serverless platforms (e.g., OpenFaaS), this paper proposes the first unified DevSecOps pipeline tailored for open-source Serverless environments. Our approach integrates static code analysis, dynamic behavior monitoring, secret scanning, and resource-constraint policies to deliver lightweight, fault-closed security checks, deeply embedded across all CI/CD stages. Unlike existing ad hoc security practices, our work provides the first systematic, low-overhead defense spanning the entire Serverless development lifecycle—effectively mitigating common threats such as injection attacks and hardcoded secrets. Evaluation on 20 real-world Serverless functions demonstrates 95% precision and 91% recall in vulnerability detection, with negligible runtime overhead.

0 citationsRead paper

A Data-Centric Approach for Safe and Secure Large Language Models against Threatening and Toxic Content

Apr 19, 2025

To address toxicity, bias, and jailbreaking risks inherent in large language model (LLM) outputs, this paper proposes BART-Corrective—a data-driven, model-agnostic post-hoc toxicity correction framework. Unlike approaches requiring architectural modifications, prompt engineering, or fine-tuning, BART-Corrective employs a lightweight, pre-trained BART-based correction model to purify LLM generations in real time. It introduces a “data-centric” post-processing paradigm, integrating multi-dimensional toxicity detection (e.g., Perspective API) and enabling plug-and-play compatibility across diverse LLMs. Extensive evaluation on four major LLMs—GPT-4, PaLM2, Mistral, and Gemma—demonstrates an average 28% reduction in toxicity and up to a 23% decrease in jailbreak success rate. The framework significantly enhances safety, generalizability, and deployment flexibility without compromising generation quality or inference efficiency.

0 citationsRead paper