Research & Papers
New AI model generates 45-minute lip-synced video from one photo and runs in real time
LPM 1.0 is a new AI model that generates 45-minute lip-synced videos in real time from a single photograph, with...
The Decoder
Research & Papers
Latest AI research
Research & PapersGoogle
A Step-by-Step Coding Tutorial on NVIDIA PhysicsNeMo: Darcy Flow, FNOs, PINNs, Surrogate Models, and Inference Benchmarking
A tutorial on implementing NVIDIA PhysicsNeMo for physics-informed machine learning, demonstrating how to build and...
MarkTechPost
Research & Papers
China has erased the US lead in AI, Stanford HAI’s 2026 AI index reveals
Stanford HAI's 2026 AI Index Report reveals that China has erased the US lead in artificial intelligence technology,...
SiliconANGLE
Research & PapersMeta
Meta AI and KAUST Researchers Propose Neural Computers That Fold Computation, Memory, and I/O Into One Learned Model
Meta AI and KAUST researchers have proposed Neural Computers (NCs), a novel architecture where neural networks function...
MarkTechPost
Research & Papers
A Coding Implementation of MolmoAct for Depth-Aware Spatial Reasoning, Visual Trajectory Tracing, and Robotic Action Prediction
This tutorial provides a practical implementation guide for MolmoAct, an action-reasoning model that performs...
MarkTechPost
Research & Papers
Researchers define what counts as a world model and text-to-video generators do not
An international research team introduces OpenWorldLib to standardize world model research and establish clear...
The Decoder
Research & Papers
Researchers from MIT, NVIDIA, and Zhejiang University Propose TriAttention: A KV Cache Compression Method That Matches Full Attention at 2.5× Higher Throughput
Researchers from MIT, NVIDIA, and Zhejiang University developed TriAttention, a KV cache compression method that...
MarkTechPost
Research & Papers
AI models would rather guess than ask for help, researchers find
Researchers tested 22 multimodal language models using ProactiveBench and found that almost none ask for help when...
The Decoder
Research & Papers
How Knowledge Distillation Compresses Ensemble Intelligence into a Single Deployable AI Model
Knowledge Distillation is a technique that compresses the intelligence of multiple ensemble AI models into a single,...
MarkTechPost
Research & Papers
Alibaba’s Tongyi Lab Releases VimRAG: a Multimodal RAG Framework that Uses a Memory Graph to Navigate Massive Visual Contexts
Alibaba's Tongyi Lab has released VimRAG, a multimodal RAG framework designed to handle visual data in...
MarkTechPost
Research & Papers
A Coding Guide to Markerless 3D Human Kinematics with Pose2Sim, RTMPose, and OpenSim
A tutorial demonstrating how to build and execute a complete markerless 3D human kinematics pipeline using Pose2Sim,...
MarkTechPost
Research & Papers
NVIDIA Releases AITune: An Open-Source Inference Toolkit That Automatically Finds the Fastest Inference Backend for Any PyTorch Model
NVIDIA has released AITune, an open-source inference toolkit that automatically selects the fastest inference backend...
MarkTechPost
Research & Papers
Five AI Compute Architectures Every Engineer Should Know: CPUs, GPUs, TPUs, NPUs, and LPUs Compared
The article examines five different AI compute architectures (CPUs, GPUs, TPUs, NPUs, and LPUs), explaining how modern...
MarkTechPost
Research & Papers
New technique makes AI models leaner and faster while they’re still learning
Researchers have developed a new technique using control theory to reduce unnecessary complexity in AI models during...
MIT News AI
Research & Papers
Sigmoid vs ReLU Activation Functions: The Inference Cost of Losing Geometric Context
This article analyzes the geometric properties of sigmoid versus ReLU activation functions in deep neural networks,...
MarkTechPost
Research & Papers
Databricks co-founder wins prestigious ACM award, says ‘AGI is here already’
Matei Zaharia, Databricks co-founder, has won the top honor from the Association for Computing Machinery. He is working...
TechCrunch AI
Research & PapersMeta
Meta AI Releases EUPE: A Compact Vision Encoder Family Under 100M Parameters That Rivals Specialist Models Across Image Understanding, Dense Prediction, and VLM Tasks
Meta AI has released EUPE, a compact vision encoder family with under 100M parameters designed for edge devices that...
MarkTechPost
Research & Papers
An Implementation Guide to Running NVIDIA Transformer Engine with Mixed Precision, FP8 Checks, Benchmarking, and Fallback Execution
This tutorial provides a practical implementation guide for NVIDIA Transformer Engine, focusing on mixed-precision...
MarkTechPost
Research & Papers
Import AI 452: Scaling laws for cyberwar; rising tides of AI automation; and a puzzle over gDP forecasting
The article discusses scaling laws for cyberwarfare applications of AI, explores the rising tide of AI automation's...
Import AI