QUELL — reproducibility package (v4.0). This repository accompanies the paper “QUELL: An On-Device Benchmark of Quantized Edge LLMs versus Classical Machine Learning for IoT Intrusion Detection.” QUELL fine-tunes decoder-only LLMs (GPT-2-medium, Qwen2.5-0.5B/1.5B, Llama-3.2-1B) on network-flow records serialized as text, quantizes the Qwen2.5-1.5B detector, deploys it on an NVIDIA Jetson Orin Nano Super (GGUF/llama.cpp), and compares it with random-forest and XGBoost detectors on Edge-IIoTset, CICIoT2023 and N-BaIoT under a leakage-controlled, multi-seed protocol. Axes: detection quality, on-device latency/energy/memory, quantization retentio
Şüyun, Süleyman Burçin, Aktürk, Fatih
What's New in v2.0.0 Chanakya - chat agent [x] On-Demand Deep Biological Evaluations : Query PubMed articles and generate structured scientific syntheses covering executive findings, assay workflows, molecular/epigenetic mechanisms, and facility takeaways. [x] Dual LLM Provider Support : Compatible with both Groq (ultra-fast Llama/Mixtral/GPT-OSS models) and Google Gemini . Pure Client-Side & Zero-Exposure Privacy [x] Direct User API Keys : Users securely provide their own API keys, stored solely within their browser's localStorage . [x] Zero Server Overhead & Zero Secret Leaks : No server-side proxies, no stored user logs, and no shared API
Das, Jyotirmoy
Source code for the multimodal LLM evaluation pipeline described in "Multimodal Large Language Models for Clinical Scar Assessment: Agreement, Bias, and Reproducibility." Implements a five-stage pipeline (image ingestion, preprocessing, multi-provider API dispatch, JSON schema validation, and Excel/CSV export) used to evaluate five MLLMs (GPT-5.4, Claude Sonnet 5, Gemini 3.1 Pro, Qwen3.7-Plus, GLM-5V-Turbo) against blinded human observer scores on the POSAS Observer Scale. Ships without patient images, API keys, or run outputs; researchers supply their own image dataset and API credentials.
Abdullah, Ahmed
Corrected analysis and figure-generation package for a benchmark of six large language models, four single-model prompting strategies, and a separate GPT-5.4-assisted external-review workflow across 119 de-identified gastrointestinal oncology cases.
Yuan, Jiajia, Zhou, Shengyuan, Yang, Guangzhengao, Zhao, Jie, Xu, Lingxiao, Jin, Zhao, Cao, Yanshuo, Chen, Zifan, Li, Yanyan, Feng, Xujiao, Dong, Bin, Shen, Lin, Chen, Yang, Zhang, Li
SYDA seamlessly generates realistic synthetic test data—including structured, unstructured, PDF, and HTML—using AI and large language models. It preserves referential integrity, maintains privacy compliance, and accelerates development workflows. SYDA enables both highly regulated industries such as healthcare and banking, as well as non-regulated environments like software testing, research, and analytics, to safely simulate diverse data scenarios without exposing sensitive information.
Lingamgunta, Rama Krishna Kumar
Updated for https://ph-ai.me Public Health & AI Workshop at University of Arizona, June 8-11, 2026
Tyson L. Swetnam, Claude, Jeffrey K Gillan, Carlos Lizarraga-Celaya, Greg T. Chism, Enrique Noriega, Eric R. Scott
This dataset contains the full source code, cron job scripts, and database schema used in the study “Early detection and personalized academic support using a predictive chatbot for student success” (submitted to PeerJ Computer Science). The system integrates a high-performance LightGBM predictive model with an AI-powered WhatsApp® chatbot based on the GPT-4o-mini model to identify at-risk students and deliver early, personalized academic support.
Arévalo-Cordovilla, Felipe Emiliano, Peña, Marta
This repository contains the full implementation of GenEneCast , an energy forecasting framework that integrates statistical decomposition (Holt-Winters), deep learning (LSTM), and Generative AI (GPT-4) to deliver interpretable, accurate, and computationally efficient predictions. The model automatically adapts its configuration based on the input time series and desired forecast precision level (Low, Medium, High), balancing trade-offs between performance and energy cost. It supports walk-forward validation, hyperparameter tuning, and generates natural language reports using GPT models. Files provided : Full source code ( .py files) README.m
Fontoura, Leonardo, de Mattos Nascimento, Daniel Luiz, Vieira Neto, Julio
Comparing LLM based approaches for predicting Mg dissolution modulators. There is a Readme with brief instructions. Copyright (c) 2025 Busch et al. If you have any questions please contact us at , or . For further details refer to our publication ( https://doi.org/10.1016/j.corsci.2025.113080 ). Please cite this work if you use any of the code or ideas provided.
Busch, Matthias, Tacke, Marius, Lamaka, Sviatlana, Zheludkevich, Mikhail, Linka, Kevin, Cyron, Christian Johannes, Feiler, Christian, Aydin, Roland C.
Ferdi Iskandar Mind & Memory — Initial Archive <table width="100%"> <tr> <td width="36%" align="center" valign="middle"> <img src="https://i.postimg.cc/Y2Z6CBzh/Chat-GPT-Image-Jun-6-2026-11-43-05-PM.png" alt="dr. Ferdi Iskandar" width="250"> <br /> <sub><b>dr codieverse</b> · Proud Public Servant</sub> </td> <td width="64%" valign="middle"> This repository is designed as a living knowledge foundation that captures Dr. Ferdi Iskandar's way of thinking, decision-making process, lessons learned, vision, and values across leadership, healthcare, artificial intelligence, governance, innovation, and human-centered system building. Purpose To preser
codieverse
Retrieval-Augmented Generation (RAG) chatbot built on LangChain with a PostgreSQL + pgvector vector store and GPT-4.1-mini that answers questions about protein-structure deposition, validation, biocuration, and wwPDB policies. Uses pymupdf4llm for Markdown-preserving PDF extraction, two-stage document chunking, Maximal Marginal Relevance retrieval, a topical guardrail, and a dual-LLM architecture with separate question-condensing and response-generation models. Deployed at https://rcsb-deposit-help.rcsb.org.
Chithari, Vivek Reddy, Young, Jasmine Y., Persikova, Irina, Liang, Yuhe, Crichlow, Gregg V., Flatt, Justin W., Ghosh, Sutapa, Hudson, Brian P., Peisach, Ezra, Sekharan, Monica, Shao, Chenghua, Burley, Stephen K.
This is the first release of the codebase for: "Interpreting Biochemical Text with Language Models: A Machine Learning Framework for Reaction Extraction and Cheminformatic Validation" This release includes: SMILES conversion from chemical names using public databases Enzymatic reaction extraction from literature using GPT models Comparison and validation against the BRENDA database Generation of cheminformatics features (atom mapping, molecular fingerprints) Classification of reaction plausibility using machine learning models All Jupyter notebooks used in the study This release is archived on Zenodo with a DOI for citation.
davenlim18
