Research & Education

BullFrog AI is an advanced AI platform designed to accelerate biopharma drug discovery and clinical development. It leverages explainable AI, graph analytics, and data harmonization tools to transform raw, fragmented biomedical data into clean, AI-ready assets. The platform features bfPREP™ for data preparation and harmonization, bfLEAP® for uncovering biologically meaningful patterns and identifying patient subgroups, and an AI Decision Support system for prioritizing drug targets and optimizing trial design. BullFrog AI aims to provide reliable insights, understand disease heterogeneity, and build risk-balanced R&D portfolios, ultimately mitigating risk and elevating precision medicine.

State is an open-source machine learning model developed by the Arc Institute, designed to predict how cells respond to various perturbations. It provides functionalities to train State Transition (ST) models for genetic perturbation prediction and pretrain State Embedding (SE) models to embed and annotate new datasets. The tool supports installation via `uv` and offers a command-line interface for preprocessing data, training models, performing inference, and transforming datasets into embeddings. It also includes optional vector database capabilities for querying similar cells. State is particularly useful for researchers in biology and bioinformatics looking to analyze and predict cellular behavior under different experimental conditions, with associated repositories for model evaluation and data loading.

Chemix, Inc. offers an innovative AI-enabled power system designed to bring safety, observability, and predictability to battery technology. By leveraging Generative AI, Chemix aims to illuminate the complexities often associated with battery performance and behavior. The company is focused on providing an end-to-end solution, with more detailed product information anticipated soon. Located in Sunnyvale, CA, Chemix is dedicated to advancing battery technology through artificial intelligence, ensuring robust and reliable power systems for various applications. Their approach seeks to enhance understanding and control over battery operations, moving beyond traditional methods to offer a more transparent and manageable power solution.

Speech-enhancement is an open-source deep learning project designed for audio denoising, specifically focusing on attenuating environmental noise from speech. The system leverages spectrograms, a 2D representation of audio, to apply Convolutional Neural Network (CNN) architectures, similar to those used in image processing. It utilizes a U-Net model, a Deep Convolutional Autoencoder with symmetric skip connections, adapted to denoise spectrograms. The project supports data creation, training, and prediction modes, allowing users to prepare datasets from various sources like LibriSpeech and ESC-50, train the U-Net model, and then predict and subtract noise models from noisy audio. It provides pre-trained weights and offers a flexible framework for speech enhancement.

Lab42 is a Swiss AI Research Institute based in Davos, dedicated to developing human-level artificial intelligence (HLAI, or AGI) to address global challenges like disease, hunger, and environmental issues. Unlike profit-oriented AI companies, Lab42 adopts a public good-oriented and open research approach, collaborating with universities worldwide and organizing AI competitions to foster innovation. The institute focuses on creating entirely new AI solutions rather than incrementally improving existing ones, emphasizing trustworthy and efficient application. Lab42 believes true intelligence involves autonomously creating solutions to unknown problems, and for this, it utilizes humanoid robots equipped with powerful AI to support humans in strenuous and dangerous tasks.

LAIA, the Open Laboratory of Artificial Intelligence, is dedicated to democratizing AI by fostering interdisciplinary exploration of technological advancements and their societal impacts. The initiative focuses on generating and promoting open AI tools to build technological sovereignty. LAIA actively engages the community through various activities, including hackathons, cine-debates, and AI model training workshops. It emphasizes the use of open AI tools to enhance diversity in voices and perspectives within the AI landscape. The platform also features a podcast, "Entre la señal y el ruido," and a blog with articles on AI topics like educational integration and language models.

3DFuse is an open-source framework designed to improve 3D consistency in text-to-3D generation by integrating 3D awareness into existing 2D diffusion models. This approach enhances the robustness of score distillation-based methods, leading to more coherent and realistic 3D outputs. The tool provides an interactive Gradio application for text-to-3D and image-to-3D generation, allowing users to preview point clouds before final 3D generation to refine desired shapes. It also includes code for 3D generation and a HuggingFace Demo for easy access. 3DFuse is built upon contributions from public projects like SJC and ControlNet, making it a valuable resource for researchers and developers in the field of AI-powered 3D content creation.

Aquabyte provides an AI and computer vision-powered solution for data-driven monitoring and insight in fish farming. The system utilizes the Hammerhead camera, engineered for submerged pens, to collect over one million images daily. This data is then processed to provide accurate datasets on fish welfare, lice, behavior, and growth. The user-friendly portal transforms complex data into actionable insights, helping farmers make informed decisions to improve efficiency, sustainability, and fish health. Aquabyte's technology has evolved from basic weight estimation and lice counting to a comprehensive platform, supported by world-class customer support with industry experience.

Sony CSL (Paris) is a dedicated research laboratory committed to advancing technologies that contribute to sustainable, peaceful, and democratic societies. The lab employs cutting-edge methodologies from complexity science, data science, and artificial intelligence to investigate fundamental issues across diverse fields, including language, music, sustainability, and information. Their work aims to foster a renewed and harmonious relationship between humanity, nature, and technology, driving innovation with a strong ethical and societal focus. The research conducted at Sony CSL (Paris) seeks to address complex global challenges through interdisciplinary approaches and scientific rigor.

LTrace Geosciences is a leader in research, development, and innovation for the energy sector, transforming complex academic research into practical, high-impact commercial solutions since 2018. Their core expertise includes Artificial Intelligence, Digital Rock Physics, History Matching, and Quantitative Seismic Interpretation. Key products include GeoSlicer, a free and open-source AI digital rock platform developed in cooperation with Petrobras and Equinor, and the LTrace Inversion Suite, which streamlines the process from data preparation to result validation for estimating elastic properties. They provide comprehensive services in Quantitative Seismic Interpretation and Reservoir Characterization, utilizing a fully integrated, probabilistic workflow for robust uncertainty quantification.

Autonoma, Inc. develops AutoVerse, a high-fidelity digital twin simulation platform specifically designed for airport airside operations. This tool enables airports to model, test, and optimize various aspects of their operations, including airside layouts, turnaround sequences, and capital projects, without disrupting live operations. It also supports surge readiness planning, safety-critical scenario validation, and the integration of autonomous systems. By simulating complex scenarios, Autonoma helps airports improve efficiency, reduce delays, and enhance safety before committing real resources or breaking ground on new infrastructure.

AutoKeras is an AutoML library built on Keras, designed to simplify and automate deep learning workflows. Developed by the DATA Lab at Texas A&M University, its primary goal is to enhance the accessibility of machine learning for a broader audience. The library automates critical tasks such as hyperparameter tuning and neural architecture search, which are often time-consuming and complex. By providing an easy-to-use interface, AutoKeras allows users to quickly build and deploy deep learning models without extensive manual configuration, making advanced AI techniques more approachable for developers and researchers alike. It is compatible with Python >= 3.7 and TensorFlow >= 2.8.0.

AMD-SHARK-Studio is a web user interface designed for SHARK+IREE, a high-performance machine learning distribution. It allows developers to run machine learning models, including Stable Diffusion, on various hardware platforms such as AMD, Nvidia, and Apple devices. The tool supports both Windows and Linux/macOS environments, offering flexible installation options for stable releases or advanced developer setups. Users can execute models via a web UI or command-line interface, with features like dispatch benchmarking for performance analysis. While the project is not currently maintained in its original form, it provides a robust framework for local ML model inference and development.

AutoRCCar is an open-source project designed to create a self-driving RC car. It integrates a Raspberry Pi, Arduino, and various open-source software components to achieve autonomous navigation. The Raspberry Pi collects inputs from a camera module and an ultrasonic sensor, transmitting this data wirelessly to a computer. The computer then processes these inputs for object detection, specifically identifying stop signs and traffic lights, and for collision avoidance. A neural network model, running on the computer, makes predictions for steering based on the input images. These predictions are subsequently sent to the Arduino for controlling the RC car. The project provides detailed instructions for setting up the environment using Anaconda, calibrating the Pi Camera, collecting training data, and training the neural network model.

Manifold is an AI platform specifically designed for the life sciences sector, aiming to accelerate the journey of medicines from initial discovery to market. It offers a comprehensive suite of tools including Agent OS for automating R&D tasks, AI Cohort Explorer for building research-ready cohorts, and AI Analysis Agent for transforming natural language questions into analyses. The platform emphasizes collaboration with features like a Collaborative Workbench and Dataset Access Controls, while supporting various tools such as Python, R, and nf-core pipelines. Manifold also provides robust data management capabilities, including AI-powered data cataloging, engineering, and dashboards for multimodal biomedical data. It ensures data protection through built-in guardrails, enterprise-grade security, and compliance with standards like NIST 800-171, TX-RAMP, SOC 2 Type II, HIPAA, and GDPR.

awesome-deep-learning-single-cell-papers is an open-source repository dedicated to curating and categorizing the most recent academic papers focused on single-cell analysis utilizing deep learning techniques. The papers are organized by specific tasks, such as multimodal learning, single-cell data simulation, interpretability, and spatial-temporal transcriptomics, making it easier for researchers to navigate and find relevant studies. The repository also includes sections for pretrained models, GANs/diffusion models, and various single-cell application tools. It serves as a valuable resource for academics and researchers looking to stay updated on advancements in deep learning applications within single-cell biology.

Hemispheric offers foundational AI solutions specifically designed for decoding the brain. The company's core mission is to apply advanced artificial intelligence techniques to push the boundaries of neuroscience research. By leveraging AI, Hemispheric aims to significantly improve our understanding of the complex mechanisms and functions of the brain. Their tools are developed to serve researchers and developers working at the intersection of AI and neuroscience, providing them with powerful capabilities to analyze neural data, model brain activity, and explore new hypotheses. This specialized focus positions Hemispheric as a key player in accelerating scientific discovery within the neuroscientific community.

Basil Systems offers an AI-powered platform designed for the life sciences industry, specifically targeting pharmaceutical and MedTech organizations. It integrates commercial, regulatory, medical affairs, quality, and safety data into a single intelligence platform, providing hourly updates across global markets and AI-validated insights. With over 30 years of historical data, Basil Systems enables teams to transition from reactive to proactive strategies, significantly compressing response times. The platform supports various functions, including identifying actionable market opportunities, accelerating regulatory submissions, strengthening scientific engagement, mastering post-market surveillance, and overcoming competitive threats. It boasts over 700 million records, real-time monitoring, and rapid implementation.

Blank Bio is at the forefront of precision medicine, utilizing advanced RNA intelligence to transform drug development and clinical trials. The platform employs foundation models that integrate isoform, mutation, and expression signals from RNA, moving beyond traditional gene-level counts. This comprehensive approach allows for the identification of complex, multi-gene patterns that are often missed by conventional methods, leading to improved patient stratification, enhanced diagnostic accuracy, and more effective therapeutic design. Blank Bio's technology supports applications such as building multi-gene signatures for trial enrichment, refining disease classification, discovering new therapeutic targets, and optimizing RNA-based therapeutics. It also offers biosecurity monitoring through automated characterization of biological threats.

The Conference on Robot Learning (CoRL) is an annual international event dedicated to advancing the fields of robotics and machine learning. It serves as a vital platform for researchers, academics, and industry professionals to share cutting-edge research, discuss new advancements, and foster collaboration. CoRL 2026 is scheduled to take place in Austin, Texas, US, from November 9-12, 2026, with workshops on November 9th and the main conference from November 10-12th. The event includes calls for papers, instructions for authors, and opportunities for sponsorship, making it a key gathering for the robot learning community.

immunitoAI is a TechBio company specializing in the development of AI-generated novel antibody therapeutics. Their platform focuses on 'Design over Discovery,' creating antibodies in silico without relying on biological sources, which aims to increase efficiency and expand the antibody space beyond natural repertoires. The company emphasizes 'Drug-First Antibodies,' embedding pre-defined drug properties at the inception of the design process to minimize the exploratory phase in drug developability. Utilizing a 'Structure-First AI Platform,' their neural networks are trained on shape complementarity for precise, target-specific design, optimizing affinity and specificity, and potentially opening possibilities for previously undruggable targets. immunitoAI's vision is to make antibody therapy the norm by unleashing the full potential of antibody-based biological drugs.

The Learning Systems and Robotics Lab (LearnSysLab), formerly the Dynamic Systems Lab, is a research group led by Prof. Angela Schoellig at the Technical University of Munich and the University of Toronto Institute for Aerospace Studies. Their research is driven by the vision of seamless interaction between robotic systems and the physical world, particularly addressing challenges in unstructured, uncertain, and changing environments. They combine ideas from controls, machine learning, and optimization to develop next-generation robot algorithms that integrate a-priori information with operational data. Their work includes developing advanced robot control and learning algorithms for real-world applications and collaborating on interdisciplinary robot projects.

PaddleHelix is an advanced bio-computing platform developed by PaddlePaddle, designed to accelerate research and development in drug discovery, vaccine design, and precision medicine. It utilizes state-of-the-art machine learning approaches, particularly deep neural networks, to offer solutions for complex biological problems. Key functionalities include large-scale pre-training models for compounds and proteins, molecular property prediction, drug-target affinity prediction, and molecular generation. For vaccine design, it provides RNA design algorithms like LinearFold and LinearPartition. The platform also supports precision medicine through applications like drug-drug synergy. Recent advancements include HelixFold3.2 for significant improvements in protein-related tasks and structural quality, HelixDock for protein-ligand structure prediction, and HelixFold-Multimer for protein complex structure prediction. PaddleHelix offers both open-source code for non-commercial academic research and paid API access for high-throughput use.

Nami Energy is at the forefront of integrating cutting-edge software and artificial intelligence with sustainable energy to create innovative solutions for a cleaner future. The company focuses on smart fusion technology, transforming everyday hardware into intelligent systems through advanced software. Nami Energy emphasizes hardware excellence, adhering to European standards and family business values to ensure high-quality products. Their approach is purpose-driven, prioritizing user needs in every innovation to deliver solutions that are both effective and user-centric. By harnessing the synergy between software innovation, AI, and clean energy, Nami Energy is dedicated to redefining a sustainable and prosperous future.