ZK
Editor-in-Chief · CPO of AlpinaGPT

Zhemal Khamidun

Head of AI at Alpina Digital, CPO of AlpinaGPT (~30-person engineering & AI team). Founder of Vibe Academy & Hamidun Consulting, author of the AI channel «Готовим ИИшницу» (@jemal_hamidun).

10,676Published articles
50+Verified primary sources
6Language editions (RU/EN/ES/PT/FR/AR)
New Benchmark Tests Whether AI-Generated Images Obey the Laws of Physics
Research·Jiqizhixin (机器之心)

New Benchmark Tests Whether AI-Generated Images Obey the Laws of Physics

Oct 12, 2026·2 min
UNITAS: A 3D World Model for Robots Claims the Lead on Several Benchmarks
Robotics

UNITAS: A 3D World Model for Robots Claims the Lead on Several Benchmarks

Oct 12, 2026·2 min
HIL-UMI: Dong Hao's Team Finds a Robot's Weak Spots Without the Robot
Robotics

HIL-UMI: Dong Hao's Team Finds a Robot's Weak Spots Without the Robot

Oct 12, 2026·2 min
Mamba-based S²FI grades fruit quality from video, backed by a dataset of 6,460 clips
Robotics

Mamba-based S²FI grades fruit quality from video, backed by a dataset of 6,460 clips

Oct 12, 2026·2 min
SAMF Teaches Drones to Locate Themselves on a Satellite Map Using Thermal Imaging, No GNSS Required
Robotics·arXiv cs.CV

SAMF Teaches Drones to Locate Themselves on a Satellite Map Using Thermal Imaging, No GNSS Required

Researchers propose SAMF, a framework that matches drone thermal images to satellite maps using scene structures common to both. The code and the Thermal-UAV-Paired…

Oct 12, 2026·2 min
APEX Compresses Diffusion Latents 128x and Beats a 256-Token Model on ImageNet-512
Research·arXiv cs.CV

APEX Compresses Diffusion Latents 128x and Beats a 256-Token Model on ImageNet-512

APEX orders latent channels from structure to detail and denoises them asynchronously. On ImageNet-512 it wins at 64x and 128x compression, and the code is already…

Oct 12, 2026·2 min
SWIM saves compute: audio-video generation gains up to 30.7% over Best-of-N
Research·arXiv cs.CV

SWIM saves compute: audio-video generation gains up to 30.7% over Best-of-N

SWIM combines particle search with Metropolis-Hastings under a fixed budget: on JavisBench-Mini, +30.7% over Best-of-N on JavisDiT++ and +17.0% over EvoSearch on…

Oct 12, 2026·2 min
SAM 2.1 fine-tuned on facades and interiors: mIoU rises from 49.9% to 69.6%
LLM·arXiv cs.CV

SAM 2.1 fine-tuned on facades and interiors: mIoU rises from 49.9% to 69.6%

Researchers fine-tuned SAM 2.1 on 2679 images of 42 buildings, lifting mIoU across 11 building-element classes from 49.9% to 69.6%. The model was integrated into…

Oct 12, 2026·2 min
Top themes
01SAIL Keeps Its Target in Frame Using Audio and Text, and Beats Rivals on Two BenchmarksOct 12, 2026·2 min02Wearable Stereo Camera Estimates Food Volume With 9.1% Error, No Reference Object NeededOct 12, 2026·2 min03Researchers Compare CNN, ViT and Mamba Attention Against Human GazeOct 12, 2026·3 min04Stroop Test Shows VLMs Read the Text in an Image Before They See the ColorOct 12, 2026·2 min05ForcingBooster Halves Denoising Steps in Autoregressive Video GenerationOct 12, 2026·2 min
MemReg Outperforms Pairwise Methods and SLAM in LiDAR Point Cloud Registration
Robotics·arXiv cs.CV

MemReg Outperforms Pairwise Methods and SLAM in LiDAR Point Cloud Registration

Oct 12, 2026·2 min
Video-Language Models Confuse Car Maneuvers: LoRA and Probes Fix the Error
Robotics·arXiv cs.CV

Video-Language Models Confuse Car Maneuvers: LoRA and Probes Fix the Error

Oct 12, 2026·2 min
MIRAGE anchors video-based 3D reconstruction to LiDAR and beats COLMAP on three datasets
Robotics·arXiv cs.CV

MIRAGE anchors video-based 3D reconstruction to LiDAR and beats COLMAP on three datasets

Oct 12, 2026·2 min
OMNI: One Compact Encoder Absorbs the Knowledge of 10 Models for Pathology
Medicine·arXiv cs.CV

OMNI: One Compact Encoder Absorbs the Knowledge of 10 Models for Pathology

Oct 12, 2026·2 min
TerraAgent checks whether the AI actually looks at the image: hallucinations drop from 17.1% to 9.8%
Research·arXiv cs.CV

TerraAgent checks whether the AI actually looks at the image: hallucinations drop from 17.1% to 9.8%

A multi-agent wrapper for satellite VQA nearly halved answers given "from memory": from 46.2% to 21.7%. No fine-tuning needed; the cost is 2.6 passes of the base…

Oct 12, 2026·2 min
AVA improves multimodal ICL accuracy by 1.7 points without any training
Research·arXiv cs.CV

AVA improves multimodal ICL accuracy by 1.7 points without any training

A preprint proposes AVA, a three-stage pipeline that cleans up the examples in a frozen VLM's prompt. It gains +1.7 points on average across five benchmarks, with…

Oct 12, 2026·2 min
Coral-KV Compresses Video Model KV-Cache by 37–48% and Lifts PSNR by Up to 4.95 dB
Research·arXiv cs.CV

Coral-KV Compresses Video Model KV-Cache by 37–48% and Lifts PSNR by Up to 4.95 dB

Researchers propose Coral-KV, a KV-cache compression method for autoregressive video generation. On MAGI-1 it gains 4.95 dB PSNR over QVG-4 while using 37% less memory.

Oct 12, 2026·2 min
Black Lake: Chinese Factories Skip Digitization and Go Straight to AI Agents
Business·SCMP Tech

Black Lake: Chinese Factories Skip Digitization and Go Straight to AI Agents

Oct 12, 2026·2 min
T. Rowe Price: 55% of Hong Kong Residents Expect AI to Hit Their Retirement, but Only One in Five Feels Ready
Business·SCMP Tech

T. Rowe Price: 55% of Hong Kong Residents Expect AI to Hit Their Retirement, but Only One in Five Feels Ready

Oct 12, 2026·2 min
Genetic Optimizer Cuts Aerosol Sensor Placement Time in the UK by 60%
Research·arXiv cs.LG

Genetic Optimizer Cuts Aerosol Sensor Placement Time in the UK by 60%

Oct 12, 2026·2 min
Kronecker Product Attention Compresses the KV-Cache Up to 10x and Matches MLA on Quality
Research·arXiv cs.LG

Kronecker Product Attention Compresses the KV-Cache Up to 10x and Matches MLA on Quality

Oct 12, 2026·2 min
TimesFM 3.0 Cuts University Building Energy Forecast Error by 43%
Research·arXiv cs.LG

TimesFM 3.0 Cuts University Building Energy Forecast Error by 43%

Oct 12, 2026·2 min
Wasserstein BaryBooster Turns Diffusion Realism into Event Forecast Accuracy
Research

Wasserstein BaryBooster Turns Diffusion Realism into Event Forecast Accuracy

Oct 12, 2026·2 min
DiffusionART speeds up RL fine-tuning of video models 12x, with no EMA tuning
LLM

DiffusionART speeds up RL fine-tuning of video models 12x, with no EMA tuning

Oct 12, 2026·2 min
A Physical Transformer Built Entirely From Transistors Learns In-Context Learning in Simulation
Research

A Physical Transformer Built Entirely From Transistors Learns In-Context Learning in Simulation

Oct 12, 2026·2 min
Theory Shows Masked Diffusion on Transformers Learns from Õ(LK/ε) Samples
Research·arXiv cs.LG

Theory Shows Masked Diffusion on Transformers Learns from Õ(LK/ε) Samples

A new arXiv paper gives end-to-end guarantees for masked diffusion models: sample complexity scales as Õ(L·K_par/ε), not exponentially as V^L in sequence length.

Oct 12, 2026·2 min
SAZO speeds up LLM fine-tuning without backprop: up to 6× fewer queries to reach the same loss
Research·arXiv cs.LG

SAZO speeds up LLM fine-tuning without backprop: up to 6× fewer queries to reach the same loss

A new zeroth-order method, SAZO, learns to tell informative training examples from uninformative ones and reaches the same loss with up to 6× fewer queries than…

Oct 12, 2026·2 min