← 전체 목록

2026-10-18 AI 리서치 브리핑

최신 VLM, sLLM, on-device AI 논문과 연구 블로그를 한눈에 정리합니다. 중복 기사 방지를 위해 URL 기준으로 추적합니다.

총 18건 요약 자동 생성

VLM 업데이트

멀티모달 비전-언어 모델의 최신 논문과 리더보드 변화

Perceive, Predict, and Plan: a Unified World Action Model for Egocentric Dexterous Manipulation

Paper arXiv cs.CV (recent)

We introduce P$^{3}$-WAM, a unified world action model for egocentric dexterous manipulation that learns 3D perception, action-conditioned prediction, and goal-directed planning through three complementary modes: Perceive, which reconstructs articulated hand motion and camera trajectories from video; Predict, which generates visual futures under prescribed actions; and Plan, which jointly generates action trajectories and their anticipated visual consequences conditioned on an initial image, a language instruction, and an optional visual goal. Our model integrates a compact, structured 3D action representation into a pretrained video diffusion transformer through a mixture-of-transformers architecture with shared base weights, modality-specific adaptation, and bidirectional visual-action attention. Randomized modality visibility lets video and 3D actions serve as either conditioning context or prediction targets across the three modes. Experiments show that unified training improves all three capabilities over single-mode training, including gains in planning quality under a vision-language model evaluation. Through unified training, our model outperforms specialized baselines on 3D hand reconstruction and action-conditioned video prediction across four datasets, while producing goal-directed manipulation videos with stronger instruction adherence and action progression than video-generation baselines. Project Page: https://p3-wam.pages.dev/

원문 보기

sLLM 트렌드

경량화·효율화를 위한 스몰 LLM 연구

Waypoint-OPD: Transferring RL-Induced Policy Shifts via Gated On-Policy Distillation

Paper Hugging Face Papers

Waypoint-OPD: Transferring RL-Induced Policy Shifts via Gated On-Policy Distillation에 관한 최근 업데이트입니다. 자세한 내용은 원문 링크에서 확인할 수 있습니다.

원문 보기

Gen4P: Video Generators are 4D Perceivers

Paper arXiv cs.CV (recent)

We introduce Gen4P, a unified framework for recovering 4D geometry and motion from monocular video. We show that a pretrained video generator can be repurposed into a feed-forward 4D perceiver without iterative denoising. Building on on-demand querying, Gen4P processes clean video latents in a single encoder pass and predicts the requested 4D attributes through a lightweight cross-attention decoder. The on-demand query formulation enables flexible spatial sampling, and our proposed continuous-time queries further provide temporal flexibility. To strengthen geometric knowledge, we propose a straightforward plug-in solution by incorporating a pretrained video-depth LoRA as a frozen geometry prior, improving overall performance without adding extra trainable parameters. Gen4P achieves state-of-the-art or competitive performance on tasks of 3D point tracking, video depth and camera pose estimation. Qualitative results on in-the-wild videos illustrate its capability of tracking complex motion while maintaining the delicate structure. Together, these results demonstrate the potential of video generative pretraining for unified 4D perception. See https://gen4p.github.io for more visual results.

원문 보기

On-Device AI

디바이스 내 추론 및 엣지 최적화 동향

SoftCache: Adaptive Step Size Caching for Diffusion Transformers

Paper arXiv cs.CV (recent)

Diffusion and flow models generate a sample over iterative steps, each a full pass through a large network, and that pass dominates generation cost. Training-free caching accelerates them by reusing activations from recent evaluations instead of recomputing them. Dynamic caching methods decide when to refresh by accumulating a cheap probe of the network input into a staleness signal and spending a fresh evaluation once it crosses a threshold. They perform hard caching: the signal is a binary switch between refresh and reuse, and every step, fresh or cached, advances by the prescribed step size. We observe that the signal carries much more than this switch. How far it sits below the threshold, its headroom, predicts the error induced by reuse. We propose SoftCache, which performs soft caching: instead of a binary switch, the headroom sets how far each cached step goes. Where the headroom is large, the sampler jumps past several scheduled nodes, and as a refresh nears, its steps shrink back toward the original step. These adaptive jumps may land between the scheduled nodes, taking the sampler off its fixed grid. SoftCache requires no training and adds no per-backbone settings: on FLUX for images and on HunyuanVideo and Wan2.1 for video, at speedups from $3.5\times$ to $6.9\times$, it is ahead of every baseline on all four fidelity metrics.

원문 보기

AI 뉴스 & 리서치

기업/연구기관의 주요 발표와 블로그 업데이트

OmniCompass: A Holistic Benchmark for Omni-Modal Video Grounding and Understanding

Paper Hugging Face Papers

OmniCompass: A Holistic Benchmark for Omni-Modal Video Grounding and Understanding에 관한 최근 업데이트입니다. 자세한 내용은 원문 링크에서 확인할 수 있습니다.

원문 보기

MJev-Spatial: System One Spatial Intelligence via Native Attention Matching

Paper Hugging Face Papers

MJev-Spatial: System One Spatial Intelligence via Native Attention Matching에 관한 최근 업데이트입니다. 자세한 내용은 원문 링크에서 확인할 수 있습니다.

원문 보기

OpenGCUA: Toward a Fully Open Stack for Generalist Computer Use

Paper Hugging Face Papers

OpenGCUA: Toward a Fully Open Stack for Generalist Computer Use에 관한 최근 업데이트입니다. 자세한 내용은 원문 링크에서 확인할 수 있습니다.

원문 보기

FDE-Gym: The Last Mile of AI Delivery

Paper Hugging Face Papers

FDE-Gym: The Last Mile of AI Delivery에 관한 최근 업데이트입니다. 자세한 내용은 원문 링크에서 확인할 수 있습니다.

원문 보기

Which Tokens to Merge? Diffusion Dynamics for Efficient Image Generation

Paper arXiv cs.CV (recent)

Token merging has emerged as an effective training-free strategy for accelerating diffusion models by reducing redundant token computation while minimizing generation degradation. A key challenge is determining which tokens can be safely merged and which should be preserved for subsequent computation. Existing method addresses this by preserving prompt-aligned foreground regions, thereby protecting semantically relevant structures. However, this text-relevance criterion overlooks background objects, local details that are not explicitly mentioned in the prompt, leading to degraded structural coherence. In this work, we propose Dynamics-aware Token Merging (DaTMe), a training-free framework that grounds token merging in the underlying diffusion dynamics. Our key insight is that token importance should reflect whether a token remains unresolved during denoising and whether it carries structural information. To this end, DaTMe constructs a dynamics- aware importance map from the Tweedie denoised estimate. The map combines a temporal signal, measuring inter-step changes in the predicted clean image, with a structural signal, capturing local high-frequency content such as boundaries and fine details. Using this map, DaTMe adaptively assigns tokens into merging roles under a given compression ratio: important tokens are preserved as anchors, while tokens that are both temporally stabilized and structurally homogeneous are selected as safe merge candidates. Experiments on PixArt-a and FLUX demonstrate that DaTMe maintains visual fidelity across the entire image, highlighting diffusion dynamics as an effective token-level merging decision in efficient generation. Our code is available at https://github.com/SeungJuCha/DaTMe.

원문 보기

T2-z: Test-Time Optimization of Environment Latents for Robotic World Models

Paper arXiv cs.CV (recent)

Predicting how the world responds to an action requires more than what can be seen: robots must account for environmental factors that are neither directly observable nor explicitly specified, from physical properties such as friction and mass to human-defined environmental rules. We introduce T2-Z for test-time optimization of environment latents, which augments a robotic world model with a learnable environment latent to condition world predictions. This compact latent captures underlying environmental variation without supervision of what environmental factors exist or what values they take. During training, T2-Z alternates between two stages: a world-model optimization stage that fixes the latents and learns general action-conditioned dynamics across environments; and an environment-latent optimization stage that freezes the world model and optimizes the conditioning latents, allowing latent structure to emerge purely through world prediction. At test time, the world model remains frozen and a newly initialized environment latent is optimized from one or more test-time interactions, adapting the model to previously unseen environmental configurations. Across 7 simulated settings and 4 real-world robotic tasks, T2-Z reduces object prediction error metric from 15.20 to 7.00 and raises action success from 33.73% to 72.96%. Its emergent latents capture structured representations of individual and compounded factors across diverse visual scenes. Ablations further show that T2-Z can efficiently infer environment latents from as little as a single test-time interaction trajectory.

원문 보기

October 14, 2026 Research@ 2026: 10 ways we're amplifying human ingenuity General Science · Generative AI · Health & Bioscience

News Google Research Blog

October 14, 2026 Research@ 2026: 10 ways we're amplifying human ingenuity General Science · Generative AI · Health & Bioscience에 관한 최근 업데이트입니다. 자세한 내용은 원문 링크에서 확인할 수 있습니다.

원문 보기

October 16, 2026 Grounding AI agents in biological reality: Building blocks for autonomous wet labs General Science · Health & Bioscience · Robotics

News Google Research Blog

October 16, 2026 Grounding AI agents in biological reality: Building blocks for autonomous wet labs General Science · Health & Bioscience · Robotics에 관한 최근 업데이트입니다. 자세한 내용은 원문 링크에서 확인할 수 있습니다.

원문 보기

October 14, 2026 Analyzing technical documents with Deep Review General Science · Machine Intelligence · Natural Language Processing

News Google Research Blog

October 14, 2026 Analyzing technical documents with Deep Review General Science · Machine Intelligence · Natural Language Processing에 관한 최근 업데이트입니다. 자세한 내용은 원문 링크에서 확인할 수 있습니다.

원문 보기

October 14, 2026 Putting Earth AI to work: Geospatial agents for our planet's biggest challenges Earth AI · General Science · Machine Intelligence

News Google Research Blog

October 14, 2026 Putting Earth AI to work: Geospatial agents for our planet's biggest challenges Earth AI · General Science · Machine Intelligence에 관한 최근 업데이트입니다. 자세한 내용은 원문 링크에서 확인할 수 있습니다.

원문 보기

October 7, 2026 Does better work always mean better workers?

News Google Research Blog

October 7, 2026 Does better work always mean better workers?에 관한 최근 업데이트입니다. 자세한 내용은 원문 링크에서 확인할 수 있습니다.

원문 보기

Proof-based general-purpose computing with Magnetar

News Microsoft Research Blog

Proof-based general-purpose computing with Magnetar에 관한 최근 업데이트입니다. 자세한 내용은 원문 링크에서 확인할 수 있습니다.

원문 보기

Simulating student learning for better AI tutoring

News Microsoft Research Blog

AI 튜터링 개선을 위한 학생 학습 시뮬레이션에 대한 최신 소식입니다. 더 자세한 내용은 원문 링크를 통해 확인하실 수 있습니다.

원문 보기

Agent Lightning v1.0: A 3,500-Line Lightweight Agentic RL Framework for Training Agents with Real Harnesses

News Microsoft Research Blog

Agent Lightning v1.0은 실제 하네스를 사용하여 에이전트를 훈련시키는 3,500줄의 경량 에이전트 RL 프레임워크입니다. 이 프레임워크에 대한 최신 업데이트는 원문 링크에서 더 자세히 확인할 수 있습니다.

원문 보기

참고한 소스