Searcharxiv⌕ Search

arXiv subjects

Yuxiang Liu

Publications and source records attributed to Yuxiang Liu.

At least 19 recordsLinked to original sources

Learning to Simulate Individuals from Macro Social Signals

Large language models are increasingly used to simulate how individuals respond to new situations, yet the behavioral reasoning behind these responses is either inherited from pretraining or learned from individual-level annotations, which offer limited behavioral diversity and little supervision of the reasoning itself. We propose to learn behavioral reasoning from prediction markets, whose price trajectories record how populations respond to real-world events at scale. We introduce macro2mind, which trains a language model with GRPO using market signals. A social behavioral decomposition makes behavioral reasoning an explicit step of forecasting: the model infers representative groups of market participants, predicts how each interprets the news and updates its beliefs, reasons about their interactions, and aggregates these responses into a price. A hindsight-regret curriculum with difficulty-aware sampling focuses training on transitions where hindsight-identified groups substantially improve the forecast while prioritizing examples that remain learnable for the current policy. The learned reasoning applies to user simulation without further training. On SWM-Bench, macro2mind achieves state-of-the-art directional accuracy and correlation on Polymarket. Trained on market data, it transfers zero-shot to four user-simulation benchmarks (Humanual, OvertonBench, PRISM, and CAD) and has competitive performance among zero-shot methods. Used as a data generator, macro2mind also raises a downstream simulator's accuracy on unseen users by 15.5 points, outperforming data generated by its backbone by 13.2 points.

cs.LG↗

Constraining Dust Properties Along the Minor Axis of Andromeda

Previous studies of dust emission in M31 based primarily on Herschel data have reported inconsistent radial trends in the dust emissivity index ($β$), partly due to limited wavelength coverage and the intrinsic $T_{\rm dust}$-$β$ degeneracy. We aim to improve the constraints of $β$ determinations and obtain the spatial surface density and temperature distributions at GMC scales in M31 by extending the spectral coverage into the millimeter regime and by quantifying the impact of wavelength coverage on the $T_{\rm dust}$-$β$ degeneracy. We present new IRAM 30m NIKA2 1.15 mm observations along the minor axis of M31 at an angular resolution of 11 arcsec, corresponding to $\sim 40$ pc at the distance of M31. These data are used with Herschel and JCMT SCUBA-2 maps to perform pixel-by-pixel single-component modified blackbody fitting. Including the 850 $μ$m and 1.15 mm data helps reduce the $T_{\rm dust}$-$β$ degeneracy and yields an average emissivity index of $β\approx 1.8$, systematically lower than values derived from far-infrared-only analyses. Radial analysis shows higher $β$ values in the 5 kpc ring compared to the 10 kpc ring. The elevated $β$ values in the inner 5 kpc may be linked to changes in dust composition, potentially associated with enhanced starlight heating in this region. Millimeter wavelength coverage is important for improving $β$ constraints in M31. The lower $β$ values inferred from the 250--1150 $μ$m SED fitting suggest possible variations in dust emissivity and highlight the importance of sampling the long-wavelength SED slope in nearby galaxies.

astro-ph.GA↗

Co-Evolving Acting and Forecasting for Language Agents

Training language agents is bottlenecked by the high cost of long-horizon rollouts, yet standard agentic reinforcement learning (RL) learns only from a terminal scalar reward, leaving the supervision in intermediate steps largely unused. Prior work exploits this supervision through value targets that estimate per-step success probabilities or observation targets that fit environment response. Both are hard to construct well: (1) value targets come from the same single scalar signal for each step in the trajectory, and (2) observation targets, such as web pages and terminal logs, are long and noisy and reach only one step ahead. We propose CoPE, which adds a third kind of multi-step action targets. On successful on-policy trajectories, CoPE trains the policy backbone to generate the action sequences that follows each history. Each action is also supervised when forecast from the preceding histories, forming a temporal ensemble that the shared backbone absorbs during training, and this enables acting and forecasting to co-evolve. The policy still acts one step at a time at inference, with no overhead. Across five benchmarks (ALFWorld, ScienceWorld, WebShop, $τ^2$-bench, and AppWorld), CoPE (1) achieves stronger aggregate performance than GRPO and RL baselines with predictive auxiliary targets across simulated environments, while improving overall performance on both realistic applications; (2) delivers up to $1.9\times$ and $2.0\times$ higher sample efficiency than ECHO and GRPO, respectively, on ScienceWorld; and (3) achieves these gains with a single policy backbone, without additional models, separate pretraining, or inference overhead. The codebase is available at https://github.com/ulab-uiuc/CoPE.

cs.AI↗

RACER: Residual-Adaptive Closed-Loop Estimation for Sampling-Based Planning in Wheeled-Quadruped Racing

We present RACER, a hierarchical control framework for wheel-based quadruped racing that combines an MPPI planner with a learned residual dynamics model and a low-level RL velocity tracker. The planner augments a nominal unicycle kinematic model with a neural residual term to capture the closed-loop tracking behavior of the RL policy. To train this residual model under limited real-world data, we propose Low-Rank Residual Adaptation (LoRRA), a two-stage approach that pre-trains on large-scale simulation data for broad coverage and then fine-tunes on a small real-world dataset with a low-rank constraint. In simulation, we empirically validate our engineering choices by showing (A) Residual dynamics improve the overall performance of our pipeline by capturing the tracking error of RL velocity tracker at high-speed cornering. (B) Residual dynamics trained with both source-domain and target-domain data gives racing performance significantly better than the residual dynamics trained with only target-domain data. (C) Low-rank constraint at target-domain adaptation gives higher success rates and higher performance than full-tune and from-scratch when domain gap in ground coefficient or joint gain increases.

cs.RO↗

Orbital-engineered px,y-kagome lattice in a halogen monolayer

Multi-orbital kagome lattices with explicit orbital degrees of freedom remain largely unexplored, as most experimentally realized systems rely on complex d-electron manifolds that are approximated by isotropic single-orbital models. Here, we overcome this limitation by realizing a px,y-orbital kagome lattice through deposition of a Br monolayer on Ag(111), where orbital filtering selectively suppresses the pz channel. Scanning tunneling microscopy, angle-resolved photoemission spectroscopy, and density-functional-theory calculations reveal a large-area, highly ordered kagome structure whose band dispersions quantitatively match the anisotropic px,y tight-binding model. To extract the intrinsic manifold from the substrate background, we construct an effective H-passivated model, which uncover the intrinsic electronic structure and reveals nontrivial topological characteristics of the px,y kagome manifold driven by first-order spin-orbit coupling effect. Our work establishes Br/Ag(111) as an experimentally accessible platform for multi-orbital kagome physics, extending the kagome paradigm from the conventional d-orbital regime to an orbitally engineered topological setting.

cond-mat.mtrl-sci↗

The Layer Mystery of VLA: An Information-Theoretical Analysis of VLA Latent Interface

Vision-language-action (VLA) policies connect a pretrained vision-language backbone to an action head through a latent interface, but which backbone layers this interface should expose remains unclear. We study single-layer selection and multi-layer fusion for frozen backbones across three pretrained models and two manipulation benchmarks, LIBERO and CALVIN, with three policy-training seeds per configuration. Across three fusion mechanisms and three layer-subset strategies, 47 of 54 configurations underperform the best observed single-layer policy. Our stastical analysis further confirms that fusion's advantage is very limited. However, the best layer varies substantially across backbones and benchmarks, making layer selection consequential and exhaustive policy sweeps expensive. We further derive a reweighting equivalence between the proposed information-bottleneck objectives for action-conditioned InfoNCE and action prediction, motivating InfoNCE as a proxy for layer quality. Empirically, InfoNCE provides the most consistent positive association with policy success among four evaluated proxies. Selecting the layer with the highest InfoNCE score requires 9-33 times less GPU compute than exhaustive policy sweeps and reduces mean selection regret from 17.89 percentage points for deepest-layer selection to 3.71 points across six settings. Its mean regret is close to the 3.28-3.50 points achieved by fixed-layer heuristics optimized retrospectively using all six oracle sweeps, without requiring closed-loop evaluations during selection.

cs.AI↗

EgoTSR++: Egocentric Spatiotemporal Reasoning for Task Progress Understanding

Vision-Language Models (VLMs) have advanced rapidly in static visual understanding, yet remain unreliable when judging how an egocentric task is progressing. Given a task instruction and two visual observations, a model should determine which state is closer to the goal by analyzing task-relevant object configurations and spatial relations, rather than relying on timestamps or presentation order. This distinction is critical in manipulation, where retries, corrective actions, and temporary regressions make progress inherently non-monotonic. We introduce EgoTSR, a unified framework for diagnosing and improving order-robust task-progress understanding. First, SpatialLogic-Bench evaluates each physical state pair in both original and order-swapped presentations across short- and long-horizon settings, exposing whether a model follows task-state evidence or chronological shortcuts. Second, our data construction pipeline converts successful, approximately monotonic manipulation and first-person trajectories into bidirectional supervision; LongTag further preserves intermediate subtask structure for long-horizon comparison, while failure-aware data extend learning to regressions and recoveries. Third, a progressive CoT-to-Tag curriculum first supervises evidence-grounded interpretation of task-relevant state changes and then consolidates the comparison rule through scalable label-only training. Experiments reveal substantial input-order bias in representative VLMs. EgoTSR achieves 92.4% long-horizon accuracy with a 0.1-point forward-inverse Gap. Failure-aware supervision further improves accuracy on non-monotonic trajectories by 11.8 points and Recovery Accuracy by 11.2 points, while maintaining broad visual and spatial capabilities. These results establish goal-conditioned state comparison as an explicit formulation of egocentric spatiotemporal reasoning for task-progress understanding.

cs.CV↗

Novel Light Dark Matter Detection with Quantum Parity Detector Using Qubit Arrays

We present the design and the sensitivity reach of the Qubit-based Light Dark Matter detection experiment. We propose the novel two-chip design to reduce signal dissipation, with quantum parity measurement to enhance single-phonon detection sensitivity. We demonstrate the performance of the detector with full phonon and quasiparticle simulations. The experiment is projected to detect $\gtrsim 30$ meV energy deposition with nearly $100\%$ efficiency and high energy resolution. The sensitivity to $m_χ\gtrsim 0.01$ MeV dark matter scattering cross section is expected to be advanced by orders of magnitude for both light and heavy mediators, and similar improvements will be achieved for axion and dark photon absorption in the $0.04$-$0.2$ eV mass range.

hep-ph↗

Extremal spectral result of outerplanar graphs without $P_{3\cdot l}$

A graph $G$ is $F$-free if it does not contain $F$ as a subgraph. Let $\mathrm{spex}(n,F)$ be the maximum spectral radius over all $n$-vertex $F$-free outerplanar graphs. For integers $t\geq1$ and $l\geq2$, let $P_{t\cdot l}$ be the starlike tree with $t$ branches of length $l-1$. For sufficiently large $n$, Yin, Li, and Meng [arXiv:2504.04364v1] characterized the unique extremal graph for $\mathrm{spex}(n,P_{t\cdot l})$ when $t=1$, $t=2$, or $t\geq4$. They left the case $t=3$ open and proposed a natural candidate for the extremal graph. We show that this candidate is not extremal and determine the unique extremal graph for $\mathrm{spex}(n,P_{3\cdot l})$. For every $l\geq3$ and all sufficiently large $n$, this unique extremal graph is $K_1\vee\bigl(2P_{2l-3}\cup qP_{l-2}\cup P_r\bigr),$ where $q$ and $r$ are integers satisfying $n=2(2l-3)+q(l-2)+r+1,$ $q\geq0,$ $0\leq r<l-2.$

math.CO↗

When Does Reasoning Help in Machine Translation? A Hierarchical Analysis of LRM Reasoning Traces

Large Reasoning Models increasingly use intermediate traces for machine translation, but it remains unclear when such reasoning helps or hurts. We analyze reasoning traces across models, languages, domains, and datasets, focusing on reasoning language, length, and structure. We find that the best reasoning language is model-specific, reasoning length has a non-monotonic relationship with quality, and traces exhibit recurring functional patterns. To uncover these patterns, we introduce Hierarchical Meta-Summarization (HMS), a scalable framework that induces coarse- and fine-grained reasoning structures without predefined taxonomies. HMS reveals a shared organization--understanding/planning, translating/drafting, and refining/verifying--alongside domain-specific variation. Our results suggest that MT reasoning should be controlled in a model-aware, length-aware, and pattern-aware manner rather than uniformly encouraged.

cs.CL↗

Low Star Formation Efficiency in M31: Evidence for a Deficit of Bound Gas

The Andromeda galaxy (M31) exhibits a systematically low star formation efficiency (SFE) compared to other star-forming galaxies. Star formation is thought to be regulated by the mass of gravitationally bound gas. To investigate whether the low SFE in M31 originates from a scarcity of such bound gas, we analyze the dynamical state of its giant molecular cloud (GMC) population by estimating their virial parameters ($α_{\rm vir}$). Using a JCMT-SCUBA2 850 micron dust-continuum GMC catalog as the parent sample, we combine it with IRAM 30 m CO(1-0) data to define a working sample of 173 GMCs and derive their kinematic properties. Most analyzed GMCs exhibit high $α_{\rm vir}$, indicating that they are predominantly gravitationally unbound at the approximately 86 pc resolution of our observations. While no statistically significant global correlation is found between SFE and $α_{\rm vir}$ for the full cloud sample, the upper quantiles of the SFE distribution show progressively stronger negative trends with increasing $α_{\rm vir}$. This upper-tail trend may arise from the combined effects of the virial parameter limiting the upper range of SFE values accessible to the GMC population and feedback-driven disruption reducing the number of high-SFE, high-$α_{\rm vir}$ CO-bright clouds in the observed sample. The substantial scatter in the SFE-$α_{\rm vir}$ plane likely reflects the diverse evolutionary stages of GMCs, the temporal mismatch between gas and star formation tracers, and the limitations of a cloud-averaged virial parameter as a tracer of internal dynamical structure, making a simple one-to-one correspondence between $α_{\rm vir}$ and instantaneous SFE difficult to recover for individual GMCs.

astro-ph.GA↗

TurboT2VA: Fast Large-Scale Text-to-Video-Audio Generation via Score-Regularized Consistency Distillation

Joint text-to-video-audio generation produces synchronized visual and acoustic content, but the long sampling trajectories and heterogeneous multimodal computation of large models make inference prohibitively expensive. We present TurboT2VA, a distillation and inference framework for accelerating a 19B-parameter joint video-audio model. Large-scale T2VA distillation is challenged by modality-imbalanced optimization, the difficulty of continuous-time consistency training at scale, and the quality--diversity trade-off. TurboT2VA addresses these issues with per-modality normalization and a progressive curriculum comprising discrete consistency warm-up, continuous consistency refinement, and joint consistency--distribution matching. The curriculum first establishes a stable, diverse generation trajectory and only then introduces distribution-level refinement. On LTX-2, four-step distillation reduces generator latency from 50.52s to 2.51s at the standard evaluation resolution of 512$\times$768, achieving a 20.1$\times$ speedup while maintaining strong visual quality, audio fidelity, diversity, and video-audio synchronization. We further develop an architecture-aware inference stack that combines guarded W8A8 and fused operators, padded-text compaction, and modality-aware sparse attention while preserving dense cross-modal and text-conditioning paths. Under the high-resolution deployment setting at 1024$\times$1792, the complete stack reduces generator latency from 318.74s to 5.83s on one NVIDIA H20, achieving a 54.67$\times$ generator-only speedup. Inference code and generation demos are available at https://github.com/thu-ml/TurboDiffusion/tree/main/turbot2va.

cs.CV↗

Dense Cores in the Vicinity of an HII Region

Massive stars strongly influence their surroundings through radiative and mechanical feedback, but its effects on dense gas structures at sub-pc scales remain poorly constrained. We investigate how feedback from a newly formed massive star affects dense cores in the filamentary molecular cloud IRAS 18530+0215. We analyze ALMA Band 6 observations of 1.3 mm dust continuum and DCN, N$_2$D$^+$, and $^{13}$CS line emission, together with VLA K-band continuum and NH$_3$ observations. Dense cores are identified with astrodendro, and their temperatures, masses, velocity dispersions, and virial parameters are derived. The dynamical state of the ultra-compact H II region is examined through energy and pressure estimates. The H II region has a radius of $\sim$0.1 pc and an expansion velocity of $\sim$2.5 km s$^{-1}$, corresponding to a shell dynamical age of $\sim$0.06 Myr. DCN and $^{13}$CS cores are concentrated near the H II region, whereas N$_2$D$^+$ cores preferentially lie farther away. Core temperatures and velocity dispersions decrease with projected distance from the H II region. Virial parameters increase within the inner $\sim$0.3 pc but decline sharply beyond this scale, while core masses show no significant trend with distance. Strong star formation signatures are found at $\sim$0.2 pc, whereas more distant regions still host quiescent, cold dense cores. The compact H II region appears trapped or choked within $\sim$0.1 pc, while its feedback extends to at least $\sim$0.3 pc. Within this region, feedback enhances core velocity dispersions, gas temperatures, and virial parameters, with no evidence that it promotes the formation of more massive dense cores.

astro-ph.GA↗

Robust and sparse support vector machine via hybrid truncated loss for supervised classification

The support vector machine (SVM) is a widely used classifier, but choosing an appropriate loss function remains difficult. Convex losses such as the hinge loss and least-squares loss are sensitive to outliers, while bounded non-convex losses often lead to high computational cost. To address this, we propose a hybrid truncated loss function ($L_{\mathrm{ht}}$) that is both sparse and bounded, and build the $L_{\mathrm{ht}}$-SVM model for single-view classification. We introduce the P-stationary point and use it to establish the first-order necessary and sufficient optimality conditions. Based on these conditions, we design an alternating direction method of multipliers with a working-set strategy that reduces computational cost and achieves global convergence. We further extend $L_{\mathrm{ht}}$-SVM to multi-view learning by adding structural information and view weights, resulting in Mv$L_{\mathrm{ht}}$-SVM, which follows both the consensus and complementarity principles. Experiments on synthetic, real-world, and image datasets show that $L_{\mathrm{ht}}$-SVM achieves higher accuracy with fewer support vectors and better noise robustness than five single-view methods, while Mv$L_{\mathrm{ht}}$-SVM outperforms six multi-view methods in accuracy, precision, recall, and F1-score.

cs.LG↗

State-Conditional Adversarial Learning: An Off-Policy Visual Domain Transfer Method for End-to-End Imitation Learning

We study visual domain transfer for end-to-end imitation learning in a realistic and challenging setting where target-domain data are strictly off-policy, expert-free, and scarce. We first provide a theoretical analysis showing that the target-domain imitation loss can be upper bounded by the source-domain loss plus a state-conditional latent KL divergence between source and target observation models. Guided by this result, we propose State- Conditional Adversarial Learning, an off-policy adversarial framework that aligns latent distributions conditioned on system state using a discriminator-based estimator of the conditional KL term. Experiments on visually diverse autonomous driving environments built on the BARC-CARLA simulator demonstrate that SCAL achieves robust transfer and strong sample efficiency.

cs.RO↗

GLIDE: Graph-guided Leap Inference for Diffusion Estimation of Spatio-Temporal Point Processes

Spatio-temporal point processes (STPPs) provide a principled framework for modeling asynchronous events in continuous time and space. Recent diffusion-based approaches offer a flexible alternative to deterministic prediction by modeling complex conditional distributions, but their application to STPPs remains challenging: reverse sampling from pure noise is costly, and weak structural constraints in sparse spatial domains can lead to poorly localized probability mass. We propose \textbf{GLIDE} (Graph-guided Leap Inference for Diffusion Estimation), a conditional diffusion framework for next-event modeling in STPPs. GLIDE organizes historical events into a multi-scale historical graph and encodes temporal evolution and spatial topology through a dual-stream architecture, yielding a structured conditioning context for a dual-branch diffusion denoiser. It further introduces a prior-guided leap inference mechanism, in which a lightweight mean predictor provides a deterministic anchor and the reverse process starts from an intermediate diffusion step instead of from pure Gaussian noise. Experiments on multiple real-world datasets show that GLIDE improves both distribution fitting and next-event prediction, with the largest gains appearing on the spatial side. The results also indicate that prior-guided leap inference substantially reduces reverse-sampling cost while preserving the stochastic generation capability of diffusion models.

cs.LG↗

Auto-Dreamer: Learning Offline Memory Consolidation for Language Agents

Language agents increasingly operate over streams of related tasks, yet existing memory systems struggle to convert accumulated experience into reusable knowledge. Retrieval-augmented and structured memory methods record per-session observations effectively, but often couple acquisition and consolidation into a single online process, leaving the agent without a global view across sessions to discover recurring patterns, abstract shared procedures, or prune redundant entries. Inspired by complementary learning systems theory, we propose Auto-Dreamer, a learned offline consolidator for language-agent memory. Auto-Dreamer decouples fast per-session memory acquisition from slow cross-session consolidation. Given a selected working region of a typed memory bank, the consolidator treats the region as read-only evidence, performs bounded tool-use to inspect entries and provenance-linked source trajectories, and synthesizes a fresh compact replacement set that abstracts across sessions and supersedes the original region. We train Auto-Dreamer via GRPO, using end-to-end agent performance as the reward signal to learn how to consolidate memories acquired through fast online experience. Trained on ScienceWorld trajectories alone, Auto-Dreamer outperforms fixed, RL-trained, and prompted memory baselines on ScienceWorld by 7 points while using an active memory bank 12$\times$ smaller than the strongest baseline, and continues to lead on held-out ALFWorld and WebArena without retraining -- using 6$\times$ less memory than the strongest baseline on ALFWorld.

cs.CL↗

VLM-VPI: A Vision-Language Reasoning Framework for Improving Automated Vehicle-Pedestrian Interactions

Autonomous driving systems often infer pedestrian yielding behavior from geometric and kinematic cues alone, limiting their ability to reason about visual scene context and age-dependent behavioral variability. This limitation can produce delayed interventions in safety-critical encounters and unnecessary braking in benign interactions. This work introduces Vision-Language Model-based Vehicle-Pedestrian Interaction (VLM-VPI), a multimodal reasoning framework for pedestrian intent understanding and yielding-aware control in autonomous driving. The system combines three components: a multimodal perception layer that captures visual and kinematic observations, a reasoning layer that uses Qwen3-VL 8B for visual scene understanding and GPT-OSS 20B for few-shot intent reasoning, and a tiered safety controller that applies age-specific braking margins for children, adults, and seniors. In 112 CARLA scenarios, VLM-VPI achieves 92.3% intent classification accuracy, outperforming a rule-based baseline (78.4%), supervised trajectory models (73.5-82.4%), and a zero-shot LLM configuration (88.4%). Validation on 24 real-world PIE scenarios yields 87.5% accuracy, indicating functional sim-to-real transferability. Across 200 simulation cases, VLM-VPI reduces the false-alarm rate from 7.4% to 2.8% and mean intersection traversal time from 13.5 s to 11.8 s. Conflict occurrences decrease from 124 to 33, while mean minimum time-to-collision improves from 1.92 s to 4.47 s. Demographic-adaptive control further reduces conflicts by 60% for children and 54.5% for seniors compared with uniform control. These results show that an explicit vision-language reasoning layer can improve both safety and efficiency by linking pedestrian intent, demographic context, and vehicle control decisions.

eess.SY↗