Research & Education

awesome-machine-learning-in-compilers is a comprehensive, curated list of research papers, datasets, and tools dedicated to the application of machine learning in compilers and program optimization. This GitHub repository serves as an invaluable resource for researchers, academics, and practitioners looking to explore and advance the field. It categorizes papers into key areas such as Survey, Iterative Compilation and Compiler Option Tuning, Instruction-level Optimisation, Parallelism Mapping and Task Scheduling, Languages and Compilation, Auto-tuning and Design Space Exploration, Code Size Reduction, Cost and Performance Models, Domain-specific Optimisation, Learning Program Representation, ML for Compilers and Systems Optimisation, and Memory/Cache Modelling/Analysis. Additionally, it provides links to relevant books, talks, tutorials, software, benchmarks, and datasets, making it a central hub for anyone interested in the synergy between machine learning and compiler technology.

Awesome-ml-privacy-attacks is a comprehensive, open-source repository dedicated to cataloging academic papers focused on privacy attacks against machine learning models. This resource is invaluable for researchers, academics, and security professionals seeking to understand and mitigate vulnerabilities in AI systems. The curated list covers various attack types, including membership inference, reconstruction, property inference, and model extraction. Where available, the repository also provides links to the authors' code implementations, enabling practical exploration and replication of the research. It serves as a central hub for staying updated on the evolving landscape of ML privacy and security.

awesome-attention-mechanism-in-cv is an open-source GitHub repository providing a curated list of attention mechanisms and plug-and-play modules specifically for computer vision applications. This resource is designed to assist researchers and developers by offering a comprehensive collection of relevant papers, their publication links, and associated GitHub repositories. The list covers various categories including Attention Mechanisms, Dynamic Networks, Plug and Play Modules, and Vision Transformers. It aims to provide a quick reference for understanding and implementing different attention-based techniques, although it acknowledges that not all modules may be included due to the vastness of the field. Users are encouraged to contribute suggestions and improvements to enhance the list's completeness.

awesome-automl-papers is a comprehensive, curated list of resources dedicated to Automated Machine Learning (AutoML). This open-source project compiles a wide array of materials including academic papers, insightful articles, practical tutorials, informative slides, and relevant projects. It serves as an invaluable resource for anyone looking to understand or stay abreast of the rapidly evolving AutoML landscape. The repository covers key areas such as Automated Data Clean, Automated Feature Engineering, Hyperparameter Optimization, Meta-Learning, and Neural Architecture Search. It also provides an overview of various AutoML approaches and their applications, making it a central hub for both newcomers and experienced professionals in the field.

The TAILOR Network of Excellence Centres on Trustworthy AI is an EU project dedicated to establishing the scientific foundations for Trustworthy AI. It achieves this by integrating learning, optimization, and reasoning (LOR) to develop AI systems that are lawful, ethical, and technically and socially robust. The project, though concluded, leaves a significant legacy in European AI, including a comprehensive Handbook of Trustworthy AI and a Strategic Research and Innovation Roadmap. TAILOR fostered collaboration between industry and academia through Theme Development Workshops and various funding initiatives, aiming to advance AI research and ensure its responsible development.

Bag_of_Tricks_for_Image_Classification_with_Convolutional_Neural_Networks is an open-source project that provides a platform for experimenting with and implementing various training tricks to improve the accuracy of image classification using Convolutional Neural Networks (CNNs). Inspired by the paper "Bag of Tricks for Image Classification with Convolutional Neural Networks," this repository tests popular techniques such as Xavier initialization, warmup training, no bias decay, label smoothing, random erasing, linear scaling learning rate, and cosine learning rate decay. It uses the CUB_200_2011 dataset and a VGG16 network for experiments, offering a practical resource for researchers and developers looking to optimize their CNN models.

The Distributed AI Research Institute (DAIR) is an independent, globally distributed organization of academics, activists, and engineers dedicated to community-rooted research. DAIR aims to cut through AI hype, exposing the real harms of AI systems while imagining and building alternative technological futures centered on care, safety, and possibility. Their work is grounded in lived experience, ensuring research addresses real problems and benefits everyone. DAIR's research areas include using data for change, identifying AI harms, envisioning alternative tech futures, and developing governance frameworks for AI systems. They prioritize comprehensive, principled research and invest in the well-being of their researchers.

JarvisIR is an AI-powered image restoration tool designed to enhance and improve the quality of digital images. Users can upload images suffering from common problems such as blur, darkness, or noise. The tool intelligently analyzes the uploaded image, identifies the specific issues, and then recommends and applies the most suitable restoration algorithms to address them. The result is a processed, restored version of the image, aiming to elevate its overall perception and clarity. While the current live website indicates a runtime error, the intended functionality is to provide an intelligent solution for various image restoration needs.

deep-learning-time-series is a comprehensive, open-source resource for anyone interested in applying deep learning to time series forecasting. This GitHub repository provides a curated list of state-of-the-art papers, code implementations, and experimental results. It covers a wide range of topics, from classic methods to deep learning approaches, and includes information on competitions and theoretical resources. The repository is continuously updated with new research from conferences like AAAI and ICLR, offering insights into various techniques such as Autoformers, N-BEATS, and attention-based models. It serves as an invaluable tool for researchers and practitioners looking to explore, implement, and stay current with the latest developments in time series forecasting using deep learning.

detrex is an open-source research platform designed for Transformer-based detection algorithms, built upon Detectron2 and borrowing design principles from MMDetection and DETR. It serves as a comprehensive toolbox for object detection, segmentation, pose estimation, and various visual recognition tasks. The platform emphasizes a modular design, allowing users to easily construct customized models, and offers strong baselines for Transformer-based detection models with optimized hyper-parameters. Key features include a LazyConfig System for flexible configuration and a lightweight training engine. detrex also provides extensive documentation, a model zoo, and supports a wide array of methods like DETR, Deformable-DETR, DINO, and MaskDINO, making it a valuable resource for researchers and developers in the field.

World Labs is a spatial intelligence company focused on developing advanced AI models capable of perceiving, generating, reasoning, and interacting with the 3D world. Their primary product, Marble, allows users to create spatially consistent, high-fidelity, and persistent 3D environments from multimodal inputs like text, images, videos, or 360 panoramas. Users can precisely control 3D layouts, interactively edit specific elements, and expand or combine worlds to build larger, more immersive experiences. The platform supports versatile outputs, enabling downloads and exports in various 2D and 3D formats for seamless integration into existing workflows in fields such as art, film, gaming, AR/VR, robotics, and architecture.

finetune-transformer-lm provides the code and model for the research paper "Improving Language Understanding by Generative Pre-Training." This open-source project is designed for researchers and developers interested in replicating and experimenting with the generative pre-training techniques described in the paper. Specifically, it includes an implementation for the ROCStories Cloze Test, allowing users to run experiments and analyze results. While the code is provided as-is with no expected updates, it serves as a valuable resource for understanding the foundational concepts of generative pre-training and language understanding models. The repository also notes that the code is currently non-deterministic due to various GPU operations, with a median accuracy slightly lower than the paper's reported single run.

HLearn is a high-performance machine learning library developed in Haskell, designed to offer both speed comparable to low-level languages like C/C++ and flexibility akin to high-level languages such as Python. It distinguishes itself by leveraging functional programming principles and the SubHask library for fast numerical computations. The library's design is deeply rooted in abstract algebra, utilizing concepts like homomorphisms, monoids, and Abelian groups to enable features such as parallel batch training, online training, fast cross-validation, and weighted data points. HLearn also incorporates a unique History monad for debugging optimization procedures without runtime overhead. While it's a research project aiming for an optimal interface, its current focus is on foundational algebraic structures rather than a broad range of popular machine learning techniques.

KENLG-Reading is a comprehensive repository dedicated to knowledge-enhanced text generation, offering a meticulously curated reading list, tutorials, papers, codes, datasets, and leaderboards. It serves as an invaluable resource for researchers and practitioners in the field, providing a survey published in ACM Computing Survey'22. The repository is actively maintained and updated, ensuring access to the latest advancements and high-citation papers. It covers various aspects of text generation, including basic NLG papers, pretrained language models, controllable generation methods, and knowledge-enhanced techniques using knowledge bases, knowledge graphs, and grounded text.

InstructIR is an AI tool designed for high-quality image restoration, guided by human-written instructions. Developed by mv-lab and presented at ECCV 2024, this model offers an all-in-one solution for various image degradation problems. Users can input an image along with natural language prompts to restore images from multiple degradation types, such as denoising, deraining, deblurring, dehazing, and low-light image enhancement. InstructIR has demonstrated state-of-the-art results, improving over previous all-in-one restoration methods. The project also provides a novel benchmark dataset for future research in text-guided image restoration. It offers a Hugging Face demo and Google Colab tutorial for easy access and experimentation, making advanced image restoration accessible through intuitive text commands.

Interpretable-ml-book is an open-source resource offering a detailed guide to interpretable machine learning. This book, available for free online, as an ebook, or in paperback, addresses the critical need for transparency in machine learning decisions. It introduces techniques to make black-box models more understandable, covering algorithms for simple interpretable models and methods for analyzing complex models. The resource is designed for machine learning practitioners, data scientists, statisticians, and stakeholders who need to trust and explain AI decisions. It aims to foster a future where machines can clearly articulate their reasoning, making the transition into an algorithmic age more human-centric.

Leaderboard serves as a robust and comprehensive benchmarking platform specifically designed for Automatic Speech Recognition (ASR). It addresses the critical need for measurable performance in ASR systems by offering three core components: a TestSet Zoo, a Model Zoo, and a Benchmarking Pipeline. The TestSet Zoo includes a wide range of academic and SpeechIO-curated datasets covering various speech recognition tasks and scenarios in both English and Chinese. The Model Zoo comprises a collection of commercial APIs and open-source models for comparison. The platform provides a simple and well-specified pipeline for data preparation, recognition, post-processing, and error rate evaluation, enabling researchers and developers to easily benchmark, reproduce, and examine ASR systems.

Machine-Learning-for-Algorithmic-Trading-Second-Edition_Original is an open-source GitHub repository accompanying the "Machine Learning for Algorithmic Trading, Second Edition" book published by Packt. This comprehensive resource aims to show how machine learning can add value to algorithmic trading strategies in a practical yet comprehensive way. It covers a broad range of ML techniques, from linear regression to deep reinforcement learning, and demonstrates how to build, backtest, and evaluate trading strategies driven by model predictions. The repository contains over 150 notebooks that put the book's concepts, algorithms, and use cases into action, providing numerous examples for working with market, fundamental, and alternative data, training models, and designing trading strategies. It also includes applications replicating recently published research and uses the latest software versions like pandas 1.0 and TensorFlow 2.2.

MocapNET is a real-time method for estimating 3D human pose, converting 2D body joint estimations from monocular color images directly into the popular Bio Vision Hierarchy (BVH) format. Its key contributions include a novel and compact 2D pose NSRM representation, a human body orientation classifier, and an ensemble of orientation-tuned neural networks. This allows for the decomposition of the body into upper and lower kinematic hierarchies, enabling robust pose recovery even with significant occlusions. An efficient Inverse Kinematics solver refines the neural-network-based solution, ensuring 3D human pose estimations are consistent with a target person's limb sizes. MocapNET achieves a 33% accuracy improvement over its predecessor while maintaining real-time performance of 70 fps on CPU-only execution.

ml-practical-usecases offers a comprehensive database of 650 machine learning system design case studies, compiled from over 100 leading companies such as Netflix, Airbnb, and Doordash. This resource provides practical insights into how these companies leverage ML to improve their products and operational processes. The case studies cover various machine learning applications, highlighting innovative approaches and methodologies. It serves as a valuable learning resource for understanding real-world ML implementations and system design principles. The repository acknowledges Evidently AI for the original compilation, making it a community-driven resource for the machine learning ecosystem.

Chrono Civilizations offers an interactive historical atlas, allowing users to explore 5,500 years of world history from 3500 BC to 2024 AD. This educational tool visualizes 1,477 historical events and over 2,700 dynamic territory borders across 10 major civilizations, including China, Greece-Rome, Egypt, India, and the Islamic world. Users can watch empires rise and fall in real-time by sliding through an interactive timeline, observing dynasty boundary changes and the coexistence of different empires like Tang Dynasty China and the Roman Empire. It features a bilingual Chinese/English interface and highlights cross-civilization interaction routes such as the Silk Road.

365-Days-Computer-Vision-Learning-Linkedin-Post is an open-source GitHub repository curated by Ashish Patel, offering a comprehensive, day-by-day learning journey through various computer vision concepts and models. Each entry in the repository corresponds to a LinkedIn post, providing a concise overview and a link to further resources on topics ranging from EfficientDet and YOLO Series to Vision Transformers, GANs, and advanced segmentation techniques. This resource is ideal for individuals looking to deepen their understanding of computer vision through a structured, accessible format, leveraging the power of community learning and readily available information.

Search and Detect (CLIP/OWL-ViT) is an AI tool hosted on Hugging Face Spaces, designed for advanced image search and object detection capabilities. Users can input a text query to locate images that contain particular objects and then highlight those objects within the images. The tool leverages the power of CLIP for image search and OWL-ViT for precise object detection. This makes it a valuable resource for researchers, developers, and anyone needing to test and refine AI models related to computer vision. The platform is accessible via a web interface, offering a straightforward way to interact with these sophisticated AI models.

PyMARL is a Python-based, open-source framework developed by WhiRL for deep multi-agent reinforcement learning. It provides implementations of several prominent algorithms, including QMIX for monotonic value function factorisation, COMA for counterfactual multi-agent policy gradients, VDN for value-decomposition networks, IQL for independent Q-learning, and QTRAN for learning to factorize with transformation. The framework is built using PyTorch and integrates with SMAC (StarCraft Multi-Agent Challenge) as its environment, specifically using SC2.4.6.2.69232 for the results in the SMAC paper. PyMARL supports saving and loading trained models, as well as watching StarCraft II replays, making it a comprehensive tool for researchers and developers in the multi-agent RL domain.