What if an AI could learn from its own memory like a human, getting smarter with every task? Researchers from East China Normal University, Shanghai AI Lab, and others present MIA: the Memory Intelligence Agent. It uses a "Manager-Planner-Executor" team. The Manager stores
@jiqizhixin
-
Cross-Architecture Distillation Recipe for Mamba Models
By
–
Attention to Mamba: A Recipe for Cross-Architecture Distillation Paper: https://
arxiv.org/abs/2604.14191 -
Mamba Models Match Transformer Performance Without Attention
By
–
Can you get a Mamba model to perform like a Transformer without adding Attention? Researchers from Apple, MILA, and Flat Iron Institute (including Abhinav Moudgil and Ningyuan Huang) have a breakthrough answer. They introduce a two-step distillation recipe: first, they convert
-
AI Model Self-Corrects Visual Grounding Errors With Confidence Scoring
By
–
What if a model could catch and correct its own mistakes while learning? Researchers from Peking University present a new AI method for visual grounding. Instead of just matching words to image regions, their system uses a "confidence score" to flag its unreliable guesses. It
-
KMLP: Hybrid AI Model for Web-Scale Tabular Data
By
–
What if a single AI model could automatically learn from billions of messy data rows, eliminating the need for manual feature engineering? Researchers from Zhejiang University and Ant Group present KMLP, a new hybrid architecture for web-scale tabular data. It uses a clever
-
Adam’s Law: Boost LLM Performance Through Common Text Rephrasing
By
–
What if you could boost any AI's performance just by rephrasing your prompts? Researchers from FaceMind & CUHK propose "Adam's Law": a simple but powerful principle that more common, frequently seen text improves LLMs. Their method paraphrases inputs into more frequent
-
Image Generators Emerge as Generalist Vision Learning Models
By
–
Image Generators are Generalist Vision Learners Paper: https://
arxiv.org/abs/2604.20329
Project: https://
vision-banana.github.io -
Google Vision Banana: Instruction-Tuned Generalist Image Generator Model
By
–
Huge! Google just proved Image Generators are Generalist Vision Learners! They introduce Vision Banana, a model built by instruction-tuning a base image generator (Nano Banana Pro). Instead of using special systems for different tasks, they reframe every vision problem—like
-
Sol-RL: Train AI Image Generators 4x Faster
By
–
What if you could train AI image generators 4x faster without losing quality?
— 机器之心 JIQIZHIXIN (@jiqizhixin) 23 avril 2026
NVIDIA, HKU, and MIT researchers present Sol-RL.
They use a clever two-stage process: first, a super-fast, low-precision search creates a huge pool of image candidates. Then, they regenerate only the… pic.twitter.com/KokLk4tmr9What if you could train AI image generators 4x faster without losing quality? NVIDIA, HKU, and MIT researchers present Sol-RL. They use a clever two-stage process: first, a super-fast, low-precision search creates a huge pool of image candidates. Then, they regenerate only the
-
DeepSeek Releases Tile Kernels GPU Optimization for LLMs
By
–
DeepSeek just released Tile Kernels! Tile Kernels are optimized GPU kernels for LLM operations, built with TileLang. DeepSeek claims that “most kernels in this project approach the limits of hardware performance in terms of compute intensity and memory bandwidth. Some of them