arXiv subjects

Jian Li

Publications and source records attributed to Jian Li.

At least 73 records · Page 4Linked to original sources

arXiv2026

Searching for Electromagnetic Counterpart Candidates to GW231123

The detection of GW231123, a gravitational-wave (GW) event with exceptionally massive and rapidly spinning black holes, suggests the possible formation within an active galactic nucleus (AGN) disk, which provides a favorable environment for potentially generating an observable electromagnetic (EM) counterpart. We conduct a search for such a counterpart by crossmatching the GW localization with a comprehensive catalog of AGN flares from the Zwicky Transient Facility. Our analysis yields six plausible optical flare candidates that are spatially and temporally coincident with GW231123 and exhibit significant deviations from their AGN baseline flux. Although these candidates represent a crucial first step, their true nature remains inconclusive. Confirming any one of these flares via future observations would provide a landmark validation of the AGN formation channel and unlock the multi-messenger potential of this extraordinary merger.

astro-ph.HE↗

arXiv2026

SMART: Shot-Aware Multimodal Video Moment Retrieval with Audio-Enhanced MLLM

Video Moment Retrieval is a task in video understanding that aims to localize a specific temporal segment in an untrimmed video based on a natural language query. Despite recent progress in moment retrieval from videos using both traditional techniques and Multimodal Large Language Models (MLLM), most existing methods still rely on coarse temporal understanding and a single visual modality, limiting performance on complex videos. To address this, we introduce \textit{S}hot-aware \textit{M}ultimodal \textit{A}udio-enhanced \textit{R}etrieval of \textit{T}emporal \textit{S}egments (SMART), an MLLM-based framework that integrates audio cues and leverages shot-level temporal structure. SMART enriches multimodal representations by combining audio and visual features while applying \textbf{Shot-aware Token Compression}, which selectively retains high-information tokens within each shot to reduce redundancy and preserve fine-grained temporal details. We also refine prompt design to better utilize audio-visual cues. Evaluations on Charades-STA and QVHighlights show that SMART achieves significant improvements over state-of-the-art methods, including a 1.61\% increase in R1@0.5 and 2.59\% gain in R1@0.7 on Charades-STA.

cs.CV↗

arXiv2026

Regularized universal topological markers for Dirac systems

Topological markers provide an efficient and powerful characterization of topological features of many systems, especially when the translation symmetry is broken. Recently, a universal topological marker applicable in different symmetry classes of topological systems is proposed. However, it suffers from irregular behaviors at the boundary and its connection to other topological indexes remains elusive. In this work, we construct regularized universal topological markers that apply to Dirac systems by utilizing position operators that are compatible with periodic boundary conditions. The regularized markers eliminate the obstructive boundary irregularities successfully and give rise to the desired global topological invariants, such as the Chern number, consistently when integrated over all the lattice sites. Furthermore, the regularized form allows us to establish an explicit connection between the markers and some other known topological indices in two dimensions. For instance, it turns out to be equivalent to the Bott index in classes A, D, and C and equivalent to the spin Chern number in classes DIII and AII. We further examine the utility and stability of this marker in disordered scenarios. We find that its variance shows peaks at the phase boundaries, which promotes it as a useful indicator for detecting disorder-induced topological phase transitions.

cond-mat.other↗

arXiv2026

Flux Variations of Fast Radio Bursts and Their Persistent Radio Sources: Evidence for a Shared Progenitor

Fast radio bursts (FRBs) are millisecond-duration extragalactic radio transients, some of which are associated with compact persistent radio sources (PRSs), hinting at a physical connection. While several models have been proposed to explain PRSs and their connection to FRBs, direct observational tests remain limited. Here, we report for the first time a correlated trend between the long-term variation of the PRS flux density and the burst energetics of FRB 20190520B and FRB 20240114A, suggesting a physical coupling between the PRS and FRB activity. We further examine additional repeaters with compact PRSs and find no clear correlation between PRS luminosity and burst activity, likely due to the limited observations. These results are consistent with scenarios in which both the PRS and FRB activity may be powered by a common energy reservoir, such as the magnetic or rotational energy of a magnetar.

astro-ph.HE↗

arXiv2026

HRTFformer: A Spatially-Aware Transformer for Individual HRTF Upsampling in Immersive Audio Rendering

Individual Head-Related Transfer Functions (HRTFs) are starting to be introduced in many commercial immersive audio applications and are crucial for realistic spatial audio rendering. However, one of the main hesitations regarding their introduction is that creating individual HRTFs is impractical at scale due to the complexities of the HRTF measurement process. To mitigate this drawback, HRTF spatial upsampling has been proposed with the aim of reducing the measurements required. While prior work has seen success with different machine learning (ML) approaches, these models often struggle with long-range preservation of local spatial variation patterns across neighbouring source directions and generalization at high upsampling factors. In this paper, we propose a novel transformer-based architecture for HRTF upsampling, leveraging the attention mechanism to better capture spatial correlations across the HRTF sphere. Working in the spherical harmonic (SH) domain, our model learns to reconstruct high-resolution HRTFs from sparse input measurements with significantly improved accuracy. To enhance spatial coherence, we introduce a neighbour dissimilarity loss that promotes magnitude smoothness, yielding more realistic upsampling. We evaluate our method using both perceptual localization models and objective spectral distortion metrics. Experiments show that our model outperforms existing methods across several evaluation metrics in generating realistic, high-fidelity HRTFs.

cs.SD↗

arXiv2026

Population-Free Pareto Tracking for Sample-Efficient Multi-Policy MORL

Multi-objective reinforcement learning (MORL) is a fundamental framework for real-world decision-making problems involving multiple conflicting criteria. Existing multi-policy (MP) methods typically rely on online evolutionary frameworks that maintain large policy populations, leading to high sample complexity and excessive agent-environment interactions. To mitigate these limitations, we present Multi-policy Pareto Front Tracking (MPFT), a framework without a self-evolving population. It leverages an efficient Pareto-tracking mechanism initialized with single-objective extreme policies to trace the Pareto front, and further densifies sparse regions to achieve an accurate approximation of the full Pareto front. MPFT can be seamlessly integrated with advanced offline MORL algorithms, thereby substantially improving sample efficiency. We evaluate MPFT on six robotic control tasks with up to three objectives and three high-dimensional tasks with more than three objectives. Experimental results show that MPFT outperforms state-of-the-art baselines in terms of hypervolume and expected utility. It also significantly reduces agent-environment interactions. These results further demonstrate that MPFT serves as a general-purpose framework that can seamlessly integrate both online and offline MORL algorithms.

cs.LG↗

arXiv2026

Server-Proximal Aggregation for Federated Domain-Incremental Learning under Partial Participation: Task-Uniform Convergence and Backward Transfer

Real-world federated systems seldom operate on static data: input distributions drift while privacy rules forbid raw-data sharing. We study this setting as Federated Domain-Incremental Learning (FDIL), where (i) clients are heterogeneous, (ii) tasks arrive sequentially with shifting domains, yet (iii) the label space remains fixed. Two theoretical pillars remain missing for FDIL under realistic deployment: a guarantee of backward knowledge transfer (BKT) and a convergence rate that holds across the sequence of all tasks with partial participation. We introduce SPECIAL (Server-Proximal Efficient Continual Aggregation for Learning), a simple, memory-free FDIL algorithm that adds a single server-side ``anchor'' to vanilla FedAvg: in each round, the server nudges the uniformly sampled participated clients update toward the previous global model with a lightweight proximal term. This anchor curbs cumulative drift without replay buffers, synthetic data, or task-specific heads, keeping communication and model size unchanged. Our theory shows that SPECIAL (i) preserves earlier tasks: a BKT bound caps any increase in prior-task loss by a drift-controlled term that shrinks with more rounds, local epochs, and participating clients; and (ii) learns efficiently across all tasks: the first communication-efficient non-convex convergence rate for FDIL with partial participation, O((E/NT)^(1/2)), with E local epochs, T communication rounds, and N participated clients per round, matching single-task FedAvg while explicitly separating optimization variance from inter-task drift. Experimental results further demonstrate the effectiveness of SPECIAL.

cs.LG↗

arXiv2026

The Pseudospectral Method for the Dirac Equation with Confining Potential

We observe that solving the Dirac equation for confined potentials using the generalized pseudospectral (GPS) method leads to deteriorating convergence of energy eigenvalues and highly oscillatory in wave functions as the confinement radius decreases. It is found that this issue stems from the first-order differentiation formulation employed in GPS method. Motivated by this insight, we adopt the kinetically balanced generalized pseudospectral method, which incorporates the kinetically-balanced condition into the GPS method. Numerical results demonstrate that the mono-kinetically-balanced generalized pseu dospectral (MKB-GPS) method yields converged energy eigenvalues and generates smooth, continuous wave functions. This is the first application of the MKB-GPS method to confined potentials, and its effectiveness is validated for small confinement radii.

physics.atom-ph↗

arXiv2026

OHP-RL: Online Human Preference as Guidance in Reinforcement Learning for Robot Manipulation

While reinforcement learning (RL) enables robots to acquire skills autonomously, its real-world deployment is severely limited by inefficient and unsafe exploration. Human-in-the-loop interventions offer a practical solution, yet existing methods typically exploit these interventions as auxiliary training signals, without fully capturing the richer information they provide about when and how autonomy should be guided. Human interventions often encode relative preferences over behavior under safety and task constraints, rather than prescribing exact actions to imitate. Motivated by this perspective, we propose Online Human Preference as Guidance in Reinforcement Learning (OHP-RL), a framework that leverages human interventions as preference information to guide policy learning. OHP-RL introduces a state-dependent preference gate that adaptively regulates when and to what extent human interventions should shape policy learning. This design enables the agent to benefit from intermittent and imperfect human feedback while preserving autonomous exploration and stable policy optimization. We evaluate OHP-RL on three challenging real-world contact-rich manipulation tasks on a Franka robot. Across all tasks, OHP-RL consistently achieves strong success rates, faster convergence, and substantially lower human intervention effort than prior approaches. Moreover, the learned policies exhibit more stable and human-aligned behavior throughout training.

cs.RO↗

arXiv2026

PID-Guided Partial Alignment for Multimodal Decentralized Federated Learning

Multimodal decentralized federated learning (DFL) must support collaboration among agents that hold different modality subsets and often different model components, while operating over peer-to-peer (P2P) overlays without a coordinating server or a global network view. A key obstacle is that conventional multimodal training often relies on a single shared representation, which implicitly assumes that heterogeneous peers can exchange and aggregate the same model components over the same communication links. In multimodal DFL, this assumption breaks down: uni- and multimodal agents may push incompatible updates through shared overlays, weakening both inter-agent transfer and cross-modal interaction. We present PARSE, a server-free framework that brings partial information decomposition (PID) into multimodal DFL. Each agent splits its latent features into redundant, unique, and synergistic slices ("feature fission"), and performs slice-aware communication over modality-conditioned P2P overlays. During training, agents exchange only the slices that are semantically alignable with their neighbors, according to the modalities and model components they share ("partial alignment"). This design avoids centralized orchestration and gradient-surgery style conflict handling, while remaining compatible with standard DFL constraints and a range of P2P overlay topologies. Across multiple benchmarks and heterogeneous peer mixes, PARSE consistently outperforms task-, modality-, and hybrid-sharing multimodal DFL baselines while keeping per-link payloads bounded. Ablations on fusion choices and split ratios, together with qualitative feature analyses and overlay-topology studies, demonstrate the robustness and communication efficiency of the proposed slice-aware design.

cs.LG↗

arXiv2026

TrimCaching: Parameter-sharing Edge Caching for AI Model Downloading

Next-generation mobile networks are expected to facilitate fast AI model downloading to end users. By caching models on edge servers, mobile networks can deliver models to end users with low latency, resulting in a paradigm of edge model caching. In this paper, we develop a novel model placement framework, called parameter-sharing model caching (TrimCaching). TrimCaching exploits the key observation that a wide range of AI models, such as convolutional neural networks or large language models, can share a significant proportion of parameter blocks containing reusable knowledge, thereby improving storage efficiency. To this end, we formulate a parameter-sharing model placement problem to maximize the cache hit ratio in multi-edge wireless networks by balancing the fundamental tradeoff between storage efficiency and service latency. We show that the formulated problem is a submodular maximization problem with submodular constraints, for which no polynomial-time approximation algorithm exists. To tackle this challenge, we study an important special case, where a small fixed number of parameter blocks are shared across models, which often holds in practice. In such a case, a polynomial-time algorithm with a $\left(1-ε\right)/2$-approximation guarantee is developed. Subsequently, we address the original problem for the general case by developing a greedy algorithm. Simulation results demonstrate that the proposed TrimCaching framework significantly improves the cache hit ratio compared with state-of-the-art content caching without exploiting shared parameters in AI models.

cs.NI↗

arXiv2026

Rethinking 1-bit Optimization Leveraging Pre-trained Large Language Models

1-bit LLM quantization offers significant advantages in reducing storage and computational costs. However, existing methods typically train 1-bit LLMs from scratch, failing to fully leverage pre-trained models. This results in high training costs and notable accuracy degradation. We identify that the large gap between full precision and 1-bit representations makes naive adaptation difficult. In this paper, we introduce a consistent progressive training for both forward and backward, smoothly converting the full-precision weights into the binarized ones. Additionally, we incorporate binary-aware initialization and dual-scaling compensation to reduce the difficulty of progressive training and improve the performance. Experimental results on LLMs of various sizes demonstrate that our method outperforms existing approaches. Our results show that high-performance 1-bit LLMs can be achieved using pre-trained models, eliminating the need for expensive training from scratch.

cs.CL↗

arXiv2026

Systematic study of one-point kinetic energy density functionals for atomic nuclei

To explore the applicability of orbital-free density functional theory (OF-DFT) in nuclear physics, we perform a systematic benchmark of 36 one-point kinetic energy density functionals, which are originally developed for electron systems in condensed matter physics. It is found that the direct use of the original parameters for electron systems leads to inconsistent performance, with certain functionals exhibiting physically unacceptable asymptotic behaviors. However, through parameter re-optimization targeting nuclear densities, different mathematical forms of generalized gradient approximation (GGA) functionals converge to a consistent root-mean-square error of approximately 13 MeV. From a physical perspective, this consistent behavior signifies that the optimized semi-local GGAs have successfully captured the macroscopic, liquid-drop-like background of the nucleus, while the residual deviations appear as periodic oscillations at the magic numbers that could reflect the quantum shell effects.

nucl-th↗

arXiv2026

FutureWorld: A Live Reinforcement Learning Environment for Predictive Agents with Real-World Outcome Rewards

Live future prediction refers to the task of making predictions about real-world events before they unfold. This task is increasingly studied using large language model-based agent systems, and it is important for building agents that can continually learn from the real world. It can provide a large number of prediction questions grounded in diverse real-world events, while preventing answer leakage. To leverage the advantages of future prediction, we present FutureWorld, a live agentic reinforcement learning environment that closes the training loop between prediction, outcome realization, and parameter updates. Specifically, we modify and extend verl-tool, resulting in a new framework that we call verl-tool-future. Unlike standard reinforcement learning training frameworks that rely on immediate rewards, verl-tool-future stores prediction-time rollouts, backfills rewards after real-world outcomes become available, and then replays the completed trajectories for policy update. Across three open-source agents, successive FutureWorld training rounds lead to consistent improvements in prediction accuracy, probabilistic scoring, and calibration, demonstrating that delayed real-world outcome feedback can serve as an effective reinforcement learning signal.

cs.AI↗

arXiv2026

Temporal evolution of the periodic GeV signal from 4FGL J1913.2+0512 and analysis of the SS 433 / W50 lobes

SS 433 is a microquasar whose relativistic jets precess every ~162 days, providing a laboratory for jet-interstellar medium interactions. We present a comprehensive analysis of 16 years of Fermi Large Area Telescope data (August 2008-September 2024) of the SS 433/W50 field, using events in the 0.3-300 GeV range and employing pulsar gating to mitigate contamination from the bright nearby pulsar PSR J1907+0602. We detect the GeV source 4FGL J1913.2+0512 (TS = 45, where TS denotes the likelihood-ratio Test Statistic) with a power-law spectrum (photon index 2.61 +- 0.08) and confirm a GeV excess at the western lobe (TS = 17). The eastern lobe of SS 433 is hinted at with lower significance. One additional GeV excess, Fermi J1909.6+0552 (TS = 20; TS = 28 over 0.1-300 GeV), located outside the SS 433 / W50 system, is revealed after gating. Exposure-corrected Lomb-Scargle periodograms and precessional phase-folded light curves show a ~162-day modulation in 4FGL J1913.2+0512. This periodicity is prominent during the first 10 years of the mission (2008-2018) but disappears thereafter, with the phase-folded flux concentrated in precessional phases 0.0-0.5. Over the full 16-year dataset, the modulation remains detectable but with reduced significance, consistent with dilution by the later non-modulated epoch. These results indicate that the efficiency and/or geometry of gamma-ray production in the SS 433 environment evolves on multi-year timescales.

astro-ph.HE↗

arXiv2026

When Does $\ell_2$-Boosting Overfit Benignly? High-Dimensional Risk Asymptotics and the $\ell_1$ Implicit Bias

Benign overfitting is well-characterized in $\ell_2$ geometries, but its behavior under the $\ell_1$ implicit bias of greedy ensembles remains challenging. The analytical barrier stems from the non-linear coupling of coordinate selection thresholds, which invalidates standard spectral resolvent tools. To isolate this algorithmic bias, we characterize the high-dimensional risk of continuous-time $\ell_2$-Boosting over $p$ features and $n$ samples. By coupling the Convex Gaussian Minimax Theorem with delicate asymptotic expansions of double-sided truncated Gaussian moments, we analytically resolve the non-smooth $\ell_1$ interpolant. Under an isotropic pure-noise model, we prove that benign overfitting fails at the linear rate: greedy selection localizes noise into sparse active sets, and the excess variance decays at a logarithmic rate $Θ(σ^2/\log(p/n))$ for noise variance $σ^2$. We remark that while this localization mechanism should persist in the presence of signals, the exact signal-noise decomposition remains an open problem. For spiked-isotropic designs with $k^*$ head eigenvalues and $r_2 = p - k^*$ tail dimensions, the risk converges to zero when $r_{2} \gg n$, but only at a logarithmic rate $Θ(σ^2/\log(r_2/n))$, which is slower than the linear decay observed in $\ell_2$ geometries. To avoid this slow convergence, we analyze the non-smooth subdifferential dynamics of the boosting flow. This yields a tuning-free early stopping rule that, under a bounded $\ell_1$-path condition, recovers the Lasso basic inequality and attains the minimax-optimal empirical prediction rate for $\ell_1$-bounded signals.

cs.LG↗

arXiv2026

Batch-of-Thought: Cross-Instance Learning for Enhanced LLM Reasoning

Current Large Language Model reasoning systems process queries independently, discarding valuable cross-instance signals such as shared reasoning patterns and consistency constraints. We introduce Batch-of-Thought (BoT), a training-free method that processes related queries jointly to enable cross-instance learning. By performing comparative analysis across batches, BoT identifies high-quality reasoning templates, detects errors through consistency checks, and amortizes computational costs. We instantiate BoT within a multi-agent reflection architecture (BoT-R), where a Reflector performs joint evaluation to unlock mutual information gain unavailable in isolated processing. Experiments across three model families and six benchmarks demonstrate that BoT-R consistently improves accuracy and confidence calibration while reducing inference costs by up to 61%. Our theoretical and experimental analysis reveals when and why batch-aware reasoning benefits LLM systems. Our code is available at https://github.com/xuanyang19/BoT

cs.AI↗

arXiv2026

Substrate-engineered tunable bound states in the continuum and directional radiation in dielectric metasurfaces

Tunable bound states in the continuum (BICs) in metasurfaces offer powerful opportunities to control light-matter interactions, yet the role of out-of-plane symmetry breaking remains poorly understood. Here, we reveal a mechanism that enables tunable high-Q BICs and directional radiation through out-of-plane symmetry breaking in all-dielectric metasurfaces. A substrate-free metasurface composed of periodically arranged multilayer cylinders that support overlapping magnetic dipole and electric quadrupole resonances, yielding electric mirror and symmetry-protected BIC responses at 1550 nm. Introducing multilayer substrates breaks out-of-plane symmetry and excites guided modes. When the guided-mode wavelength matches that of the BIC and coupling to the substrate is suppressed, the BIC wavelength remains nearly invariant, while the Q factor increases with layer number. In contrast, spectral detuning and enhanced coupling lead to pronounced blueshifts and rapid Q degradation. The interplay between guided-mode matching and coupling strength thus governs whether a BIC remains robust or becomes tunable. These findings establish a general framework for BIC engineering via out-of-plane symmetry breaking and provide a versatile platform for tunable metasurfaces with potential applications in integrated optics.

physics.optics↗