Tag
The paper proposes a Decision Transformer-based approach for optimizing UAV-mounted RIS-assisted dynamic D2D communications, demonstrating cross-scenario generalization and efficient zero-shot transfer.
The paper proposes Latent Interface Training (LIT), a two-stage strategy to enhance generalization in robotics foundation models by mitigating vision-action shortcuts through pose-supervised latent interfaces.
This paper introduces topology as a distinct generalization axis for neural operators in PDE solving, using Hodge heat flow and the TopoBox-3D framework to show its influence across the Hodge spectrum beyond geometric compatibility.
This paper proposes optimized data mixing for supervised fine-tuning to enable reasoning models to generalize in-language reasoning across 60 languages, with a model achieving over 93% L2 reasoning rate on multiple benchmarks.
The article critiques the oversimplified view that robotics is merely an integration problem after AI breakthroughs, highlighting challenges in training data scarcity and generalization for vision-language-action models. It contrasts these limitations with the reliability of specialized industrial robots achieved through traditional automation.
This paper proposes UCF-Net, a cascaded fusion network that combines CLIP and DINO for improved deepfake image detection with better cross-domain generalization.
Scaling video pre-training to 120K hours boosts zero-shot success in World Action Models from 36.1% to 77.8% on real robots, enabling faster action prediction.
The study investigates cross-modal generalization in vision-language models for grammatical number agreement, demonstrating that these models can abstract beyond surface co-occurrence using visual cues.
The paper proposes Self-Specialized Teacher Distillation (SSTD), a two-stage method to improve domain-specific performance in language models while preserving general capabilities, without requiring replay data or external teachers.
A step-by-step walkthrough explaining how Reinforcement Learning from Human Feedback (RLHF) corrects bias in AI models, using a simple example where a single human preference about doctors generalizes to other professions like CEOs.
This paper introduces PATH, a framework for active curriculum refinement in reinforcement learning that leverages curriculum graph structure to improve robustness and generalization across diverse environments.
AutoSaddler is a method for automatically patching agent harnesses from execution traces, ensuring updates generalize by testing on a held-out set to prevent regressions.
This paper proposes an analytical framework for estimating the optimal early stopping time in neural network training using Rademacher complexity with L1-norm, avoiding probabilistic assumptions and applicable to both linear and nonlinear models.
MA-VLA is a unified framework for multi-arm robot collaboration that decomposes cooperative behavior into atomic prompts and uses training-time permutations (Arm Shuffle) to enable generalization to unseen coordination patterns.
This paper investigates AI learning and conceptual transfer in the Game of Hidden Rules, focusing on reinforcement learning with Transformer-based A2C framework, rule difficulty analysis, transfer learning, and generalization.
The tweet praises the LFM2.5-2.6B model for its strong performance in on-device evaluations, highlighting its generalization capabilities beyond agentic tasks in partnership with Artificial Analysis for testing on mobile devices.
The paper proposes dynamic context scheduling to enhance generalization in contextual reinforcement learning, showing improved performance in out-of-distribution and in-distribution regimes across simulation environments.
A robot played ping pong with 2016 Olympic champion Ding Ning, demonstrating the ability to generalize to different paddle orientations and positions.
The paper proposes OTTA-DGAD, a method for online test-time adaptation in dynamic graph anomaly detection that uses dynamic prototypes and memory buffers to handle unseen target domains without retraining.
This paper establishes tight generalization bounds for multi-dimensional hyperparameter tuning in data-driven algorithm design, using real algebraic geometry and a multi-regime lower-bound framework to resolve theoretical gaps.