Tech Times on MSN

VideoChat3 beats GPT-5 on video grounding: Open-source, full training stack released

VideoChat3, a 4-billion-parameter open-source video AI model from Nanjing University, outperforms GPT-5 and Gemini 2.5 Flash ...

The US$58.9 Billion Humanoid Robot Opportunity Starts with Actuators, Sensors, and Batteries | Says DataM Intelligence

How precision actuators, tactile sensors, advanced battery systems, and scalable manufacturing are emerging as the most ...
Tech Times

Inkling Ships: Murati’s Lab Puts Largest US Open-Weight AI on Hugging Face

Thinking Machines Lab Inkling launches as the largest US-built open-weight AI model — a 975-billion-parameter multimodal ...
XDA Developers on MSN

6 local LLMs I've used that prove they're not just smaller versions of cloud models

These local LLMs are changing the game in lots of fun ways.
GitHub

Penguin-VL: Exploring the Efficiency Limits of VLM with LLM-based Vision Encoders

Penguin-VL is a compact vision-language model family built to study how far multimodal efficiency can be pushed by redesigning the vision encoder, rather than only scaling data or model size.
IEEE

MoVAE: Multi-Omics Variational Auto-Encoder for Cancer Subtype Detection

Abstract: With the advent of advanced genomic data extraction methods, numerous studies have been utilized these data to identify cancer subtypes. Given the complexity of cancer subtyping and the ...
IEEE

Mitigating Backdoor Attacks in Pre-Trained Encoders via Self-Supervised Knowledge Distillation

Abstract: Pre-trained encoders in computer vision have recently received great attention from both research and industry communities. Among others, a promising paradigm is to utilize self-supervised ...