AI Agent Papers2026
Runs an open-source personal software agent across a user's devices to perform actions, keep readable provenance-backed memories, and coordinate via a self-hostable relay. Ships minimal end-to-end clients (mobile, desktop, web) under GPL-3.0; model-agnostic and designed for inspectability and device
Analyzes cross-tokenizer on-policy distillation (OPD) and shows that focusing supervision on strictly aligned token positions—or a student-selected top-16 subset of the shared vocabulary—retains most distillation gains, while adding span-level MSE supervision reduces downstream accuracy.
AI Agent Papers2026
Creates programmable, real-time interactive code-based environments by separating deterministic simulator state from a shared neural video renderer so agents perceive, interact, and iteratively evolve via distilled playbooks; introduces Adversarial Forcing to distill a geometry-conditioned renderer for responsive visual feedback.
Dahyun Chung, Siyoon Jin +6Affiliation: Hyunsung Kim, Affiliation: [4pt] † Co-corresponding authors +2
Generates synchronized egocentric video streams for multiple interacting agents in a shared environment, enforcing cross-view action consistency, shared environment memory, and consistent propagation of interaction-induced state changes — aimed at embodied AI, VR, and multi-agent vision research.
Measures when LLMs produce visible reasoning despite instructions not to, introducing three response-level metrics (Empty-Thinking Rate, Question–Pre-answer Relevance, Explicit Inference Rate) and evaluating 'no-thinking' behavior across tasks, models, and prompting interventions.
AI Agent Papers2026
Lets a pretrained multimodal LLM interpret navigation requests and orchestrate motion via tool calls for generalist robot navigation across unfamiliar scenes. Key features: an agent harness with Navigation Skills, a unified visual-point interface, task-progress tracking, and tool-based motion execution without navigation-specific fine-tuning.
Predicts identity-preserving dense pixel correspondences between image pairs that violate spatio-temporal priors (e.g., edits and reference-guided generation). Fuses generative (FLUX2) and semantic (DINOv3) foundation representations with heterogeneous supervision and teacher-guided iterative refinement to generalize beyond classical optical-flow assumptions.
Fine-tunes diffusion generative models with online reinforcement learning via Aggregation-First Reward Tilting (DiffusionART): it aggregates per-prompt rewards into a shared reward-improving direction and adds a trust-region restoration update to replace EMA, improving training stability and GPU efficiency for image and video generation.
Trains a three-stage system that enables humanoid robots to visually perceive and reflexively catch thrown boxes using an egocentric RGB-D encoder, a compact latent object-dynamics representation, and a whole-body RL controller; reports 85.7% simulation success and 65%/40% real-world success at 2.5/5 m.
AI Agent Papers2026
Jiahao Wang, Hongjin Qian +3Beijing Academy of Artificial Intelligence, Institute of Automation, Chinese Academy of Sciences +1
Measures whether coding agents can build agents that truly satisfy natural-language requirements via AgentSWE: a 25-task benchmark spanning Creation, Editing, and Optimization. Builders develop against few illustrative cases; acceptance runs frozen submissions on held-out cases and requirement-derived rubrics, exposing overfitting, hollow runtimes, and unsupported completion claims.
Xiaohan Qin, Xiangdong Zhang +8School of AI, Shanghai Jiao Tong University, Shanghai Innovation Institute +3
Enables multi-head, depth-wise residual retrieval in Transformers by expanding a single residual stream into multiple context-aware streams using causal depthwise convolutions. Improves pretraining convergence and downstream benchmark scores with minimal overhead, yielding notable iso-loss compute speedups in large MoE models.
Bingzhe Zhang, Kai Li +1Department of Computer Science and Technology, Institute for AI, BNRist, Tsinghua University, Beijing 100084, China, IDG/McGovern Institute for Brain Research, Tsinghua University, Beijing 100084, China +1
Adapts frequency-band boundaries to each audio mixture under a fixed band-token budget to improve source separation. Predicts content-conditioned boundary scores, selects bands with hard Top-K, pools variable-width bands into tokens, and decodes full-resolution complex masks—showing large SI-SDR gains on DnR v2, MUSDB18-HQ and EchoSet.