AllLLMRoboticsStartupsResearchEthicsBusinessHardwareAutonomousMedicineCreative
Topic

#Research

Curated coverage on «Research»: releases, research, benchmarks, open-source.

1209 articles-55% vs prev. week146 this week
Subscribe
RELATED:#LLM#Business#Hardware#Creative#Startups#Ethics#Autonomous#Robotics#Medicine
TopoGPU Computes Persistent Homology 53x Faster on Average
Research·arXiv cs.CV

TopoGPU Computes Persistent Homology 53x Faster on Average

Oct 9, 2026·2 min
GHARP animates a 3D head in real time: 13x faster on A100
Research

GHARP animates a 3D head in real time: 13x faster on A100

Oct 9, 2026·2 min
SAVER Teaches Video Models to Reason From a Few Frames, Using 1,250 Examples
Research

SAVER Teaches Video Models to Reason From a Few Frames, Using 1,250 Examples

Oct 9, 2026·2 min
Distilled Diffusion Models Learn to Forget Unwanted Concepts via DPO
Research

Distilled Diffusion Models Learn to Forget Unwanted Concepts via DPO

Oct 9, 2026·2 min
SPLIT-RL Teaches VLMs to See and Reason Separately: +1.4–6.1 pp Over GRPO
Research·arXiv cs.CV

SPLIT-RL Teaches VLMs to See and Reason Separately: +1.4–6.1 pp Over GRPO

SPLIT-RL, a new post-training scheme, splits vision-language model training into perception and reasoning phases and beats GRPO by 1.4–6.1 points on Qwen3-VL and…

Oct 9, 2026·2 min
Flow matching: velocity scaling cuts FID from 28.0 to 12.2 on ImageNet-256
Research·arXiv cs.CV

Flow matching: velocity scaling cuts FID from 28.0 to 12.2 on ImageNet-256

An arXiv paper attributes the velocity-scaling effect in flow matching to a population time lag, not to underestimated velocity. FID on ImageNet-256 fell from 28.0…

Oct 9, 2026·2 min
VICO Trains VLMs on Rewritable Images: Up to +5.0% on New Tasks
Research·arXiv cs.CV

VICO Trains VLMs on Rewritable Images: Up to +5.0% on New Tasks

The VICO framework jointly trains a VLM actor and an image editor: task difficulty adapts to the model, and no annotation is needed. Out-of-domain gains reach +5.0%.

Oct 9, 2026·2 min
DOGS Beats 9 Baseline Methods at Logo Generation Across Two Renderers
Research·arXiv cs.CV

DOGS Beats 9 Baseline Methods at Logo Generation Across Two Renderers

DOGS drops prompt rewriting: a GFlowNet samples a logo design graph and penalizes similarity to existing marks. The authors compared it with nine baseline methods.

Oct 9, 2026·2 min
Top themes
01MESSENGER cuts 3D motion forecasting error by 71.6% on nuScenesOct 9, 2026·2 min02LinSlot Teaches a Neural Network to Break a Scene Into Objects and Their Attributes Without LabelsOct 9, 2026·2 min03OcuFlow Removes Eyeglass Glare in One Step and Wins 67.32% of Votes in a Blind TestOct 9, 2026·3 min04Researchers release VR180 dataset shot on Blackmagic cameras: 1211 stereo samplesOct 9, 2026·2 min05SLVR Adds 14.2 Points on BLINK Relation to Qwen2.5-VL-7B Without Text CoTOct 9, 2026·2 min
Researchers Prove Why Adversarial Training Makes Saliency Maps Sparse
Research·arXiv cs.LG

Researchers Prove Why Adversarial Training Makes Saliency Maps Sparse

Oct 9, 2026·2 min
Distilling ECG into PPG Memorized Identity: Fixed-Effects More Than Doubled the State Signal
Research·arXiv cs.LG

Distilling ECG into PPG Memorized Identity: Fixed-Effects More Than Doubled the State Signal

Oct 9, 2026·3 min
Subliminal learning passed a backdoor through numbers: 23.5% of replies in French
Research·arXiv cs.LG

Subliminal learning passed a backdoor through numbers: 23.5% of replies in French

Oct 9, 2026·2 min
PXtal Beats Baseline Models on Six Benchmarks, Four of Them Zero-Shot
Research·arXiv cs.LG

PXtal Beats Baseline Models on Six Benchmarks, Four of Them Zero-Shot

Oct 9, 2026·2 min
Nullify Erases Knowledge From LLMs Without Retraining, Using Null-Space Vectors
Research·arXiv cs.LG

Nullify Erases Knowledge From LLMs Without Retraining, Using Null-Space Vectors

Nullify makes an LLM "forget" private data at inference time: the weights stay unchanged, and forgetting quality on TOFU and MUSE matches baseline methods.

Oct 9, 2026·2 min
Multimodal Alzheimer's AI model reaches 0.935 AUROC on 2649 patients
Research·arXiv cs.LG

Multimodal Alzheimer's AI model reaches 0.935 AUROC on 2649 patients

A new framework combines MRI, clinical transformers and ODE-GRU. On ADNI, OASIS-2 and MIRIAD, diagnostic AUROC is 0.935, but on the external OASIS-3 cohort it drops…

Oct 9, 2026·2 min
CARing teaches LLMs to predict diagnoses: R@30 hits 46.52% on MIMIC-IV
Research·arXiv cs.LG

CARing teaches LLMs to predict diagnoses: R@30 hits 46.52% on MIMIC-IV

CARing's authors encode diagnoses as Semantic IDs and reward the model for covering distinct diagnoses. In reasoning mode, R@30 is 46.04% on MIMIC-III and 46.52% on MIMIC-IV.

Oct 9, 2026·2 min
D-SLR Replaces Truncated SVD With Three Decompositions and No Tuning
Research·arXiv cs.LG

D-SLR Replaces Truncated SVD With Three Decompositions and No Tuning

Oct 9, 2026·2 min
Kolmogorov-Arnold Networks Match MLP Performance With 40% Fewer Samples
Research·arXiv cs.LG

Kolmogorov-Arnold Networks Match MLP Performance With 40% Fewer Samples

Oct 9, 2026·2 min
IKNO for Navier–Stokes Stays Accurate Under Rotation, With 5.6x Fewer Parameters Than TFN
Research·arXiv cs.LG

IKNO for Navier–Stokes Stays Accurate Under Rotation, With 5.6x Fewer Parameters Than TFN

Oct 9, 2026·2 min
LoopOPD Teaches Looped Language Models to Improve Through Their Own Loops
Research·arXiv cs.LG

LoopOPD Teaches Looped Language Models to Improve Through Their Own Loops

Oct 9, 2026·2 min
MoE Router Telemetry Leaks Fine-Tuning Data: Attack Gains 2.7–9.4 pp in Accuracy
Research·arXiv cs.LG

MoE Router Telemetry Leaks Fine-Tuning Data: Attack Gains 2.7–9.4 pp in Accuracy

Oct 9, 2026·3 min
SPERA Trained on 80,000 Hours of EEG From 29,048 People and Tested on 9 Tasks
Research

SPERA Trained on 80,000 Hours of EEG From 29,048 People and Tested on 9 Tasks

Oct 9, 2026·2 min
80 Synthetic Dialogues Are Enough to Train Activation Probes for Gemma-3-27B, but Not Everywhere
Research

80 Synthetic Dialogues Are Enough to Train Activation Probes for Gemma-3-27B, but Not Everywhere

Oct 9, 2026·2 min
Temporal Transformer for the CAN Bus Catches Anomalies Without Sharing Raw Data
Research

Temporal Transformer for the CAN Bus Catches Anomalies Without Sharing Raw Data

Oct 9, 2026·2 min
Freezing the Decoder in KV-Cache Compression Saves 3x Memory but Doesn't Improve Accuracy
Research·arXiv cs.LG

Freezing the Decoder in KV-Cache Compression Saves 3x Memory but Doesn't Improve Accuracy

An arXiv paper finds that when a KV cache is compressed with SVD, fine-tuning only the encoder matches the alternatives on quality at a third of the trainable…

Oct 9, 2026·2 min
SFT-as-Context Restores General Skills in Fine-Tuned LLMs, Within 2.2 Points
Research·arXiv cs.CL

SFT-as-Context Restores General Skills in Fine-Tuned LLMs, Within 2.2 Points

A training-free method has a base model use its SFT version's answer as context. Across 19 model pairs, general-skill losses average no more than 2.2 percentage…

Oct 9, 2026·3 min
Related topics
LLM 1628Business 1117Robotics 174Hardware 100Autonomous 17Startups 9Medicine 8Ethics 5