TorchAO: A PyTorch-Native Shortcut To Smaller, Faster Models TorchAO is PyTorch's native toolkit for model efficiency: it unifies post-training quantization (PTQ), quantization-aware training (QAT), float8 (FP8) training, and structured sparsity in one coherent... deep learning FP8 model efficiency open source PyTorch QAT quantization sparsity TorchAO
Microsoft’s Analog Optical Computer: The Next Leap for AI and High-Performance Computing As the demands of artificial intelligence and high-performance computing push traditional digital electronics to their limits, a new approach is needed. Microsoft's analog optical computer represents ... AI hardware analog computing deep learning energy efficiency future technology Microsoft optical computing
Deep Learning Resurrects Ancient Antibiotics: Mining Extinct Proteomes for Modern Medicine In a groundbreaking fusion of paleobiology and artificial intelligence, researchers from the University of Pennsylvania have pioneered a revolutionary approach to combat the escalating global antibiot... antibiotic discovery antimicrobial resistance artificial intelligence computational biology deep learning drug discovery extinct organisms machine learning molecular de-extinction paleogenomics
Revolutionizing Genome Annotation: The Power of SegmentNT Deciphering the human genome, with its 3 billion nucleotides, remains one of biology’s greatest challenges. Precisely mapping genes and regulatory elements is crucial for understanding gene expression... bioinformatics deep learning foundational models genome annotation genomics machine learning multi-species SegmentNT
AI-Powered DeepSomatic Sets New Standard in Tumor Variant Detection Genetic mutations drive cancer, but pinpointing true variants amid sequencing errors has been a challenge for researchers. DeepSomatic is an AI tool developed by Google Research and collaborators that... AI in healthcare bioinformatics cancer genomics deep learning open source tools precision medicine variant detection
Skala: The Deep Learning Breakthrough Accelerating Computational Chemistry What if you could forecast chemical reactions and design new materials with laboratory-level precision, all without setting foot in a lab? Skala, a revolutionary deep-learning-based exchange-correlati... AI for Science Azure AI computational chemistry deep learning density functional theory Microsoft Research molecular simulation
How AI Is Used For Shadow Detection in Fusion Reactors Fusion reactors operate at extreme temperatures, making the identification of vulnerable surfaces essential for safe and efficient operation. A cutting-edge artificial intelligence tool, HEAT-ML, deve... AI deep learning fusion energy magnetic shadows plasma physics PPPL reactor design SPARC
SimpleFold: When Less Is More in Protein Folding Protein folding is the process of predicting a protein's three-dimensional atomic structure from its amino acid sequence which has been revolutionized by deep learning breakthroughs like AlphaFold2 an... CAMEO22 CASP14 deep learning ensemble generation flow matching generative modeling protein folding transformer
How Foundation Models Are Transforming Scientific Discovery Foundation models such as large language models have made waves in language processing and computer vision, but their leap into scientific disciplines is both exciting and complex. Scientific fields d... data efficiency deep learning foundation models machine learning physical constraints scientific discovery time series forecasting uncertainty quantification
Unlocking New AI Potential with Deep Researcher and Test-Time Diffusion Cutting-edge artificial intelligence (AI) models excel at many tasks, but their performance often depends on how well they can adapt to new, unseen data. Deep Researcher with Test-Time Diffusion is a ... adaptation AI research deep learning generalization machine learning reasoning test-time diffusion
Scaling Laws Unveiled: The Mathematical Blueprint Behind Smarter, Cheaper AI Models Training a state-of-the-art language model can cost tens of millions of dollars and months of computation time, not to mention the expertise needed for development. Yet until now, researchers have bee... AI Chinchilla deep learning GPT-3 language models machine learning neural networks optimization principal component analysis scaling laws statistical methodology
Neural Networks Are Transforming 3D Rendering: Inside Microsoft's RenderFormer 3D rendering powers our most captivating digital experiences, from blockbuster movies to cutting-edge virtual reality. Traditionally, this field has relied on physics-based methods to recreate the int... 3D rendering AI research computer graphics deep learning machine learning neural networks RenderFormer SIGGRAPH