Research & Education

Arsenale Bioyards is a pioneering platform focused on making industrial-scale biomanufacturing economically viable. It achieves this through a proprietary system that spans from lab-scale organism design to large-scale industrial production. The platform integrates hardware, data, and AI into a seamless system, controlling every step of the production value chain. Key components include THE PICCOLO micro-scale reactors for high-throughput precision fermentation, DESIGN@SCALE, an AI-powered software for optimizing bioreactor runs, and THE BIOYARD, housing Goliath reactors exceeding 50,000 liters for industrial production. This end-to-end integration ensures precision, scalability, and continuous learning, fostering an ecosystem where components interact harmoniously to accelerate time-to-market and reduce risk.

golearn is a comprehensive machine learning library designed for the Go programming language, emphasizing both simplicity and customizability. It offers a 'batteries included' approach, providing a wide range of functionalities for machine learning tasks. Users can load data as Instances, perform matrix-like operations, and pass them to various estimators. The library implements the scikit-learn interface of Fit/Predict, allowing for easy swapping of estimators during trial and error. Additionally, golearn includes helper functions for data management, such as cross-validation and train-test splitting. It supports various algorithms including KNN, linear models, neural networks, and decision trees, making it suitable for diverse machine learning applications.

pytorch-attention offers a robust PyTorch implementation of various cutting-edge deep learning models, including a wide array of attention mechanisms, vision transformers, MLP-like models, and convolutional neural networks. This open-source codebase is designed for researchers and engineers to easily experiment with and integrate advanced architectures into their projects. It features implementations of models like Squeeze-and-Excitation Attention, ViT, ResNet, and MLP-Mixer, complete with code examples for quick setup and testing. The repository is modular and extensible, making it a valuable resource for anyone working on computer vision and deep learning tasks, providing a foundation for both academic research and practical application development.

PINNs-Torch is a PyTorch-based implementation of Physics-Informed Neural Networks (PINNs), designed to accelerate scientific computing tasks. A key differentiator is its integration of CUDA Graphs and JIT Compilers (TorchScript), which can boost performance by up to nine times compared to earlier TensorFlow v1 implementations. The package is open-source and provides a robust framework for researchers and developers to build and experiment with PINNs. It includes examples for various problems, such as the Navier-Stokes PDE, and offers flexible installation options for both users and contributors. The tool is ideal for those looking to leverage the power of PyTorch for physics-informed machine learning, with a focus on speed and usability.

RTDL (Research on Tabular Deep Learning) is a comprehensive, open-source GitHub repository dedicated to advancing the field of deep learning for tabular data. It serves as a valuable resource for researchers and practitioners by curating a collection of academic papers and associated software packages. While the original `rtdl` Python package is deprecated, the repository itself remains active, pointing users to updated and more efficient packages like `rtdl_revisiting_models` and `rtdl_num_embeddings` for implementing models such as MLP, ResNet, and FT-Transformer. The project aims to provide up-to-date research and practical implementations, allowing users to stay informed on the latest advancements and apply deep learning techniques to tabular datasets.

wincnn is a Python module specifically designed to generate minimal Winograd convolution algorithms, which are crucial for optimizing convolutional neural networks. This tool implements the algorithms proposed in the research paper "Fast Algorithms for Convolutional Neural Networks" by Lavin and Gray (CVPR 2016). It provides symbolic computation capabilities, ensuring exact results for the transforms. Users can compute transforms for various F(m,r) configurations, including examples like F(2,3), F(4,3), and F(6,3), and also generate algorithms for linear convolution. The module requires Python 3.8 or higher and SymPy 1.9 or higher for its operation, making it a valuable resource for developers and researchers working on neural network optimization.

Tata Consultancy Services (TCS) is a global leader in IT services, consulting, and business solutions, dedicated to building Perpetually Adaptive Enterprises. They leverage technology to catalyze business transformation and help organizations evolve to thrive in a constantly changing world. TCS offers a wide range of services across various industries, including cutting-edge solutions in AI, cloud security, and digital transformation. Their offerings include programs like 'My First AI Job,' reports on 'Manufacturing Cyber Threats,' and platforms such as 'Rapid Outcome AI' with NVIDIA and the 'Gemini Experience Center' for Physical AI adoption. TCS is recognized as a leader in AI services by IDC and Everest Group, demonstrating deep expertise and the ability to deliver AI at scale.

Lotus-2 Depth is an AI-powered tool designed for depth estimation, providing an official demo of the Lotus-2 model. Users can upload a photo, and the application will compute a detailed geometric prediction, such as a depth map or a surface-normal map, which reveals the 3D structure of the scene. The result is returned as an image, making it useful for various applications requiring depth information. This tool is particularly valuable for computer vision research, 3D reconstruction, and other fields where understanding the spatial relationships within an image is crucial.

Lotus-2 Normal is an AI-powered tool available as a Hugging Face Space, designed to generate depth or surface-normal maps from uploaded images. Utilizing the advanced Lotus-2 image model, it processes your input and presents the resulting map alongside the original picture. The interactive slider allows for easy comparison and visualization of the generated output. This tool serves as an official demonstration of the Lotus-2 model, making it valuable for researchers, developers, and anyone interested in computer vision applications requiring precise surface normal information or depth estimation from images.

Multitask Text and Chemistry T5 is an AI tool designed for chemistry and text-based tasks, allowing users to generate text or molecular structures from input prompts. It offers capabilities for various tasks, including predicting chemical reactions and describing actions. This tool is particularly useful for researchers and scientists who work with chemical data and require advanced text analysis or molecular structure generation. Its versatility makes it a valuable asset for exploring chemical properties and reactions through natural language processing.

iReason, LLC is a research and development company focused on delivering end-to-end AI solutions, emphasizing human-centered intelligence. Their services span from initial research to full deployment, ensuring reliability and trustworthiness within the data science community. iReason is committed to advancing beyond state-of-the-art AI, offering strategic design and deployment support. Key proprietary products include OpenBrain, a framework for developing language-specific intelligent voice bots using advanced NLP, speech processing, and knowledge representation. Another innovative product is HYPO, a novel, non-invasive embedded device for detecting hypertension based solely on ECG signals, aiming to replace traditional blood pressure measurement devices.

OmniumAI is an AI and data science company specializing in bioinformatics and chemoinformatics. It offers consulting, software development, and training services, primarily assisting biotechnological companies with complex data analysis, mining, and integration challenges. The platform aims to decode life and leverage natural processes using advanced machine learning techniques, providing solutions for scientific research and development. Its focus on these specialized fields makes it a valuable resource for organizations looking to apply AI to biological and chemical data.

NTv3 is an AI-powered tool that provides foundation models specifically designed for long-range genomics research. Hosted on Hugging Face, it offers a convenient hub for users to access ready-to-run PyTorch notebooks. These notebooks facilitate various genomic tasks, including inference, fine-tuning, interpretation, and sequence generation. Researchers can input DNA/RNA sequences or training data to leverage the models' capabilities for advanced genomic analysis. The platform is developed by InstaDeepAI, making cutting-edge AI models accessible for scientific computing in the genomics domain.

Chat01.ai, previously known as OpenAI01.net, provides users with free and unlimited access to OpenAI and ChatGPT models, positioning itself as a robust alternative to ChatGPT Pro. The platform offers a selection of models, including GPT-5.5-thinking, GPT-5.4-thinking, GPT-5.3, and GPT-4.5, with advanced 'pro' models like GPT-5.5-pro, GPT-5.4-pro, and GPT-5.2-pro available on paid plans. Users can ask questions and receive answers from these models. The service operates on a credit system, offering daily free credits and permanent credits through paid plans or invites. It aims to assist users with various queries, leveraging advanced AI capabilities.

Latitudo 40 is an innovative platform that leverages artificial intelligence and high-resolution satellite imagery to provide comprehensive geospatial insights. It continuously monitors urban environments, real estate, and critical infrastructure, delivering actionable data on land surface temperature, vegetation cover, and urban heat islands. This technology empowers cities, investors, banks, and insurance companies to proactively manage climate impacts, optimize decisions for enhanced resilience, and promote sustainability. The platform offers products like EarthDataInsight for urban planning and EarthDataPlace, a marketplace for precision satellite data and geospatial insights, supporting various markets including climate resilience, urban planning, agritech, and insurtech.

Tierra Biosciences, operating under the name Revr, is an AI-driven protein discovery engine designed to revolutionize protein-based therapeutics. The platform focuses on high-performance protein design, offering rapid development, optimization, and synthesis capabilities. By leveraging artificial intelligence, Revr aims to eliminate the guesswork traditionally associated with protein engineering, providing a more efficient and grounded approach to biological reality. This tool is geared towards accelerating the discovery and development of protein-based solutions, making it a valuable asset for professionals in the biotechnology and pharmaceutical sectors.

PhyCV is the first Physics-inspired Computer Vision Python library developed by Jalali-Lab at UCLA. It introduces a new class of computer vision algorithms that simulate the propagation of light through physical mediums with diffractive properties, followed by coherent detection. Unlike traditional empirical algorithms, PhyCV leverages physical laws as blueprints, making these algorithms potentially implementable in real physical devices for fast and efficient computation. The library currently includes Phase-Stretch Transform (PST) for edge and texture detection, Phase-Stretch Adaptive Gradient-field Extractor (PAGE) for directional edge detection, and Vision Enhancement via Virtual diffraction and coherent Detection (VEViD) for low-light and color enhancement. Both CPU and GPU versions are available for each algorithm, with GPU versions depending on PyTorch and torchvision.

PipeCNN is an OpenCL-based FPGA Accelerator specifically designed for large-scale Convolutional Neural Networks (CNNs). It leverages High Level Synthesis (HLS) tools to facilitate the design and implementation of customized circuits on FPGAs, significantly speeding up the hardware development cycle compared to traditional RTL-based methodologies. The project provides a generic, yet efficient, OpenCL-based CNN accelerator that is scalable in both performance and hardware resources, making it suitable for various FPGA platforms. PipeCNN supports both Intel OpenCL SDK and Xilinx Vitis based FPGA design flows and includes a ModelZoo with pre-quantized models for networks like VGG-16 and ResNet-50. While the performance may not match the latest state-of-the-art designs, PipeCNN serves as a complete and valuable resource for learning about Deep Learning Architecture (DLA) and experimenting with new ideas in FPGA acceleration.

breast_cancer_classifier is an open-source deep learning model designed to assist radiologists in breast cancer screening. It utilizes deep neural networks to analyze mammography images, offering both an image-only model and an image-and-heatmaps model for classification. The tool processes four standard mammography views (L-CC, R-CC, L-MLO, R-MLO) and outputs probabilities of benign and malignant findings for each breast. While the provided model is from 2019 and not clinically used at NYU Langone Health, it serves as a foundational implementation for researchers and developers interested in breast cancer classification using AI.

Holodeck is a high-fidelity simulator designed for reinforcement learning and robotics research, leveraging the power of Unreal Engine 4. It offers a robust platform with over seven rich worlds and numerous scenarios for training AI agents. The simulator supports both Linux and Windows operating systems, allowing for easy extension and modification of training scenarios. A key feature is its ability to train and control multiple agents simultaneously, providing a flexible environment for complex research. It boasts a simple, OpenAI Gym-like Python interface for ease of use and high performance, capable of simulation speeds up to 2x real-time. Holodeck can run headless or with visual feedback, catering to different research needs.

Kornia is a differentiable computer vision library built on PyTorch, designed for spatial AI applications. It offers a comprehensive suite of differentiable image processing and geometric vision algorithms, allowing users to leverage powerful batch transformations, auto-differentiation, and GPU acceleration. Key features include a wide range of image processing operators like filters, transformations, and enhancements, as well as advanced augmentation pipelines for training AI models. Kornia also provides access to pre-trained AI models for tasks such as face detection, feature matching, segmentation, and classification. The library is expanding its focus towards end-to-end vision models, with a particular emphasis on integrating state-of-the-art Vision Language Models (VLM) and Vision Language Agents (VLA). It supports multi-framework usage, including TensorFlow, JAX, and NumPy, making it a versatile tool for developers and researchers in the AI and computer vision fields.

NeuralPDE.jl is an open-source solver package designed for Scientific Machine Learning (SciML) that utilizes Physics-Informed Neural Networks (PINNs) to solve various types of differential equations, including Ordinary, Stochastic, and Partial Differential Equations (ODE, SDE, PDE). It offers a greatly increased generality compared to classical methods by leveraging neural stochastic differential equations. Key features include automated construction of physics-informed loss functions from a high-level symbolic interface, compatibility with machine learning libraries like Flux.jl and Lux.jl for GPU-powered layers, and integration with NeuralOperators.jl for mixing deep neural operators with physics-informed loss functions. The tool also supports advanced techniques such as quadrature training strategies, adaptive loss functions, and neural adapters to accelerate training, making it suitable for complex scientific simulations and data fitting.

OpenFace is a state-of-the-art, open-source toolkit designed for comprehensive facial behavior analysis. It enables real-time facial landmark detection, accurate head pose estimation, robust facial action unit recognition, and precise eye-gaze estimation. Developed by Tadas Baltrušaitis in collaboration with CMU MultiComp Lab, OpenFace is intended for computer vision and machine learning researchers, as well as the affective computing community. The tool stands out for its ability to run efficiently from a simple webcam without requiring specialized hardware, making advanced facial analysis accessible. It provides source code for both running and training models, ensuring flexibility and extensibility for research and application development.

SEAL (learning from Subgraphs, Embeddings, and Attributes for Link prediction) is a novel framework designed for link prediction. It systematically transforms the link prediction task into a subgraph classification problem. For each target link, SEAL extracts its h-hop enclosing subgraph and constructs a node information matrix, which can include structural node labels, latent embeddings, and explicit attributes. This data is then fed into a graph neural network (GNN) to classify the existence of the link, allowing the model to learn from both graph structure features and latent/explicit node features simultaneously. The framework is implemented in both MATLAB and Python, with a PyTorch Geometric version available for testing on OGB, Planetoid, and custom datasets. Notably, SEAL can achieve strong performance even without node embeddings or attributes, leveraging purely graph structures, and can function as an inductive link prediction model.