Relative Entropy Pathwise Policy Optimization
arXiv:2507.11019v4 Announce Type: replace Abstract: Score-function based methods for policy learning, such as REINFORCE and PPO, have delivered strong results in game-playing and robotics, yet their h
Knowledge catalogue
arXiv:2507.11019v4 Announce Type: replace Abstract: Score-function based methods for policy learning, such as REINFORCE and PPO, have delivered strong results in game-playing and robotics, yet their h
arXiv:2604.10423v1 Announce Type: new Abstract: Replicability requires that algorithmic conclusions remain consistent when rerun on independently drawn data. A central structural question is compositi
arXiv:2604.11278v1 Announce Type: new Abstract: In federated learning (FL), accommodating clients with diverse resource constraints remains a significant challenge. A widely adopted approach is to use
arXiv:2604.10183v1 Announce Type: cross Abstract: Wireless sensing, traditionally relying on signal processing (SP) techniques, has recently shifted toward data-driven deep learning (DL) to achieve pe
arXiv:2601.08039v2 Announce Type: replace Abstract: In this paper, we study Riemannian zeroth-order optimization in settings where the underlying Riemannian metric g is geodesically incomplete, and th
arXiv:2604.10974v1 Announce Type: new Abstract: Reinforcement learning (RL) policies often fail under dynamics that differ from training, a gap not fully addressed by domain randomization or existing
arXiv:2412.07469v3 Announce Type: replace-cross Abstract: Causal discovery is a crucial initial step in establishing causality from empirical data and background knowledge. Numerous algorithms have be
arXiv:2604.09887v1 Announce Type: new Abstract: Medical vision-language datasets are often limited in size and biased toward negative findings, as clinicians report abnormalities mostly but might omit
arXiv:2604.10404v1 Announce Type: cross Abstract: Edge-based multimodal medical monitoring requires models that balance diagnostic accuracy with severe energy constraints. Continuous acquisition of EC
arXiv:2603.09355v1 Announce Type: cross Abstract: Under the multiplicative noise scaling (MNS) condition, original Nesterov acceleration is provably sensitive to noise and may diverge when gradient no
arXiv:2604.11275v1 Announce Type: new Abstract: Spatio-temporal systems often exhibit highly heterogeneous and non-intuitive responses to localized disruptions, limiting the effectiveness of conventio
arXiv:2604.10373v1 Announce Type: cross Abstract: From adversarial robustness to multi-agent learning, many machine learning tasks can be cast as finite-sum min-max optimization or, more generally, as
arXiv:2305.09958v5 Announce Type: replace Abstract: Graph neural networks (GNNs) realize great success in graph learning but suffer from performance loss when meeting heterophily, i.e. neighboring nod
arXiv:2604.11263v1 Announce Type: cross Abstract: Accurate and efficient surrogate models for aerodynamic surface pressure fields are essential for accelerating aircraft design and analysis, yet deter
arXiv:2604.03088v3 Announce Type: replace-cross Abstract: LLM agents increasingly adopt skills as a reusable unit of composition. While skills are shared across diverse agent platforms, current system
arXiv:2604.10821v1 Announce Type: new Abstract: High-dimensional and complex discrete distributions often exhibit multimodal behavior due to inherent discontinuities, posing significant challenges for
arXiv:2604.09952v1 Announce Type: new Abstract: Many applications today use large language models for code generation; however, production systems have strict latency requirements that can be difficul
arXiv:2510.03843v4 Announce Type: replace-cross Abstract: Manually editing pasted code is a long-standing developer pain point. In internal software development at Google, we observe that code is past
arXiv:2409.02231v5 Announce Type: replace-cross Abstract: We show that large language model (LLMs) can be transformed via supervised fine-tuning (SFT) of engineered prompts into SmileyLlama for explor
arXiv:2509.12694v5 Announce Type: replace Abstract: We propose the Soft Graph Transformer (SGT), a soft-input-soft-output neural architecture designed for MIMO detection. While Maximum Likelihood (ML)
arXiv:2602.17776v2 Announce Type: replace-cross Abstract: Physics-governed models are increasingly paired with machine learning for accelerated predictions, yet most 'physics--informed' formulations t
arXiv:2601.20628v3 Announce Type: replace-cross Abstract: Cluster analysis relates to the task of assigning objects into groups which ideally present some desirable characteristics. When a cluster str
arXiv:2604.03753v2 Announce Type: replace-cross Abstract: Modern advanced driver assistance systems (ADAS) rely on deep neural networks (DNNs) for perception and planning. Since DNNs' parameters resid
arXiv:2604.09745v1 Announce Type: cross Abstract: We derive a closed-form geometric functional for kernel dynamics on finite graphs by applying the Maximum Caliber (MaxCal) variational principle to th
arXiv:2405.03063v2 Announce Type: replace-cross Abstract: We propose a generalized debiased Lasso estimator based on a stability principle. When a single column of the design matrix is perturbed, the
arXiv:2604.11479v1 Announce Type: new Abstract: As global political tensions rise and the anticipation of additional tariffs from the United States on international trade increases, the issues of econ
arXiv:2604.10371v1 Announce Type: new Abstract: This study proposes Structural Gating and Effect-aligned Discovery for Temporal Causal Discovery (SGED-TCD), a novel and general framework for lag-resol
arXiv:2604.09759v1 Announce Type: cross Abstract: Transformers achieve state-of-the-art performance in natural language processing, vision, and scientific computing, but demand high computation and me
arXiv:2501.07773v3 Announce Type: replace Abstract: Generative modeling of symmetric densities has a range of applications in AI for science, from drug discovery to physics simulations. The existing g
arXiv:2603.01420v2 Announce Type: replace Abstract: This work presents a finite element-guied physics-informed learning framework for multiphysics problems with coupled partial differential equations
arXiv:2604.11529v1 Announce Type: new Abstract: Foundation models have transformed natural language processing and computer vision, and a rapidly growing literature on time-series foundation models (T
arXiv:2604.10180v1 Announce Type: cross Abstract: Disaggregation maps parts of an AI workload to different types of GPUs, offering a path to utilize modern heterogeneous GPU clusters. However, existin
arXiv:2604.09560v1 Announce Type: new Abstract: Transformers, diffusion-maps, and magnetic Laplacians are usually treated as separate tools; we show they are all different regimes of a single Markov g
arXiv:2604.10272v1 Announce Type: new Abstract: We prove that in a coupled Kuramoto oscillator network at stable equilibrium, the physical phase displacement under weak output nudging is the gradient
arXiv:2604.11511v1 Announce Type: cross Abstract: When users exercise data deletion rights under the General Data Protection Regulation (GDPR) and similar regulations, mobile network operators face a
arXiv:1810.07793v2 Announce Type: replace Abstract: We introduce the Wasserstein Transform (WT), a general unsupervised framework for updating distance structures on given data sets with the purpose o
arXiv:2512.15992v2 Announce Type: replace-cross Abstract: We develop a quantitative approximation theory for shallow neural networks using tools from time-frequency analysis. Working in weighted modul
arXiv:2604.10553v1 Announce Type: new Abstract: Graph neural networks have demonstrated excellent applicability to a wide range of domains, including social networks, biological systems, recommendatio
arXiv:2604.02927v2 Announce Type: replace Abstract: Routing algorithms are crucial for efficient computer network operations, and in many settings they must be able to react to traffic bursts within m
arXiv:2604.11198v1 Announce Type: new Abstract: Accurate air traffic prediction in the terminal airspace (TA) is pivotal for proactive air traffic management (ATM). However, existing data-driven appro
arXiv:2604.10158v1 Announce Type: new Abstract: While modern transformer neural networks achieve grandmaster-level performance in chess and other reasoning tasks, their internal computation process re
arXiv:2604.10980v1 Announce Type: new Abstract: Diffusion models have become the de facto standard for modern visual generation, including well-established frameworks such as latent diffusion and flow
arXiv:2405.09866v2 Announce Type: replace-cross Abstract: In recent years, novel communication strategies have emerged to face the challenges that the increased number of connected devices and the hig
arXiv:2602.18679v2 Announce Type: replace Abstract: Large-scale foundation models for scientific machine learning adapt to physical settings unseen during training, such as zero-shot transfer between
arXiv:2604.10848v1 Announce Type: new Abstract: Sequence modelling requires determining which past tokens are causally relevant from the context and their importance: a process inherent to the attenti
arXiv:2604.10074v1 Announce Type: new Abstract: Transformer-based diffusion models have demonstrated remarkable performance at generating high-quality samples. However, our theoretical understanding o
arXiv:2511.00413v4 Announce Type: replace Abstract: Agentic large language model (LLM) training often involves multi-turn interaction trajectories that branch into multiple execution paths due to conc
arXiv:2602.11623v2 Announce Type: replace Abstract: We revisit the use of probabilistic values, which include the well-known Shapley and Banzhaf values, to rank features for explaining the local predi
arXiv:2604.11253v1 Announce Type: cross Abstract: Feature importance methods using unrestricted permutations are flawed due to extrapolation errors; such errors appear in all non-trivial variable impo
arXiv:2602.14200v4 Announce Type: replace Abstract: Time Series Language Models (TSLMs) are emerging as unified models for reasoning over continuous signals in natural language. However, long-context
arXiv:2604.11257v1 Announce Type: new Abstract: Graph Data Prompt (GDP), which introduces specific prompts in graph data for efficiently adapting pre-trained GNNs, has become a mainstream approach to
arXiv:2604.10952v1 Announce Type: new Abstract: Selecting prototypical examples from a source distribution to represent a target data distribution is a fundamental problem in machine learning. Existin
arXiv:2604.11729v1 Announce Type: cross Abstract: General first-order methods (GFOM) are a flexible class of iterative algorithms which update a state vector by matrix-vector multiplications and entry
arXiv:2509.20587v3 Announce Type: replace-cross Abstract: We study an unsupervised domain adaptation problem where the source domain consists of subpopulations defined by the binary label Y and a bina
arXiv:2509.10853v2 Announce Type: replace-cross Abstract: Although conceptually related, variable selection and relative importance (RI) analysis have been treated quite differently in the literature.
arXiv:2604.09943v1 Announce Type: new Abstract: Reservoir computing (RC) is a computational framework known for its training efficiency, making it ideal for physical hardware implementations. However,
arXiv:2604.09558v1 Announce Type: cross Abstract: With the widening gap between compute and memory operation latencies, data movement optimizations have become increasingly important for DNN compilati
arXiv:2604.10343v1 Announce Type: new Abstract: We study the operation of community water systems, where pumps and valves must be scheduled to reliably meet water demands while minimizing energy consu
arXiv:2604.10062v1 Announce Type: new Abstract: We study reward poisoning attacks in reinforcement learning (RL), where an adversary manipulates rewards within constrained budgets to force the target
arXiv:2604.10569v1 Announce Type: new Abstract: Decision-tree ensembles are a cornerstone of predictive modeling, and SHAP is a standard framework for interpreting their predictions. Among its variant