arXiv subjects

Tiantian Zhang

Publications and source records attributed to Tiantian Zhang.

At least 37 records · Page 2Linked to original sources

arXiv2026

Power Reinforcement Post-Training of Text-to-Image Models with Super-Linear Advantage Shaping

Recently, post-training methods based on reinforcement learning, with a particular focus on Group Relative Policy Optimization (GRPO), have emerged as the robust paradigm for further advancement of text-to-image (T2I) models. However, these methods are often prone to reward hacking, wherein models exploit biases in imperfect reward functions rather than yielding genuine performance gains. In this work, we identify that normalization could lead to miscalibration and directly removing the prompt-level standard deviation term yields an optimal policy ascent direction that is linear in the advantage but still limits the separation of genuine signals from noise. To mitigate the above issues, we propose Super-Linear Advantage Shaping (SLAS) by revisiting the functional update from an information geometry perspective. By extending the Fisher-Rao information metric with advantage-dependent weighting, SLAS introduces a non-linear geometric structure that reshapes the local policy space. This design relaxes constraints along high-advantage directions to amplify informative updates, while tightening those in low-advantage regions to suppress illusory gradients. In addition, batch-level normalization is applied to stabilize training under varying reward scales. Extensive evaluations demonstrate that SLAS consistently surpasses the DanceGRPO baseline across multiple backbones and benchmarks. In particular, it yields faster training dynamics, improved out-of-domain performance on GenEval and UniGenBench++, and enhanced robustness to model scaling, while mitigating reward hacking and preserving semantic and compositional fidelity in generations.

cs.CV↗

arXiv2026

Bridging Perception and Action: A Lightweight Multimodal Meta-Planner Framework for Robust Earth Observation Agents

Autonomous Earth Observation (EO) agents are transitioning from passive perception to complex, multi-step task execution. However, current architectures that integrate planning and execution within a single model often struggle with combinatorial complexity and reasoning errors in dynamic EO scenarios. To resolve these challenges, we propose the Lightweight Multimodal Meta-Planner (LMMP) framework. LMMP incorporates a dual-awareness mechanism that grounds strategic plans in both multimodal image features and high-level task semantics. Crucially, we introduce a Meta Task Library to inject remote sensing expert knowledge directly into the workflow, which standardizes domain logic and ensures plans are physically feasible. We further implement a two-stage training pipeline, initializing the Meta-Planner via expert-distilled Supervised Fine-Tuning and refining it through Direct Preference Optimization based on execution feedback. Extensive experiments on a dataset derived from EarthBench and ThinkGeo demonstrate that LMMP significantly improves tool-calling accuracy and task success rates. Moreover, the framework exhibits strong ``plug-and-play'' versatility, consistently enhancing the performance of diverse executor backbones across previously unseen EO missions.

cs.MA↗

arXiv2026

DDO-RM: Distribution-Level Policy Improvement after Reward Learning

Recent theory suggests that reward-model-first methods can be more sample-efficient than direct policy fitting when the reward function is statistically simpler than the induced policy. We propose DDO-RM, a finite-candidate decision-optimization method that converts reward scores into an explicit target distribution. Unlike PPO-based RLHF or DPO, DDO-RM performs a KL-regularized mirror-descent update to project the policy toward a reward-improved distribution over a candidate set. Preliminary experiments on Pythia-410M show that DDO-RM outperforms DPO in pair accuracy (0.52 to 0.56) and mean margin (0.13 to 0.53). Our framework provides a principled connection between reward learning and mirror-descent policy improvement.

stat.ML↗

arXiv2026

Spin-Orbit Driven Topological Phases in Kagome Materials

Kagome materials have garnered substantial attention owing to their diverse physical phenomena, yet canonical systems such as the AV$_3$Sb$_5$ family exhibit poor $Z_{2}$-type topological properties, spurring an urgent quest for kagome platforms hosting ideal topological states. Recently, Zhou et al. proposed the kagome-type IAMX family, which exhibits distinctive ideal topological states; however, their analysis is primarily restricted to the spinless approximation. In this work, we model relativistic effects in the IAMX family, demonstrating that tuning the spin-orbit coupling (SOC) strength drives topological phase transitions and induces novel topological states, resulting in a rich phase diagram. The configuration of topological surface states evolves continuously as the SOC strength is modulated, consistent with the evolution of the topological phase transition. This suggests a viable route toward designing multi-functional topological devices. First-principles calculations performed on three specific IAMX compounds confirm that SOC governs their topological phases, in complete accord with our model analysis.

cond-mat.mtrl-sci↗

arXiv2026

Few-for-Many Personalized Federated Learning

Personalized Federated Learning (PFL) aims to train customized models for clients with highly heterogeneous data distributions while preserving data privacy. Existing approaches often rely on heuristics like clustering or model interpolation, which lack principled mechanisms for balancing heterogeneous client objectives. Serving $M$ clients with distinct data distributions is inherently a multi-objective optimization problem, where achieving optimal personalization ideally requires $M$ distinct models on the Pareto front. However, maintaining $M$ separate models poses significant scalability challenges in federated settings with hundreds or thousands of clients. To address this challenge, we reformulate PFL as a few-for-many optimization problem that maintains only $K$ shared server models ($K \ll M$) to collectively serve all $M$ clients. We prove that this framework achieves near-optimal personalization: the approximation error diminishes as $K$ increases and each client's model converges to each client's optimum as data grows. Building on this reformulation, we propose FedFew, a practical algorithm that jointly optimizes the $K$ server models through efficient gradient-based updates. Unlike clustering-based approaches that require manual client partitioning or interpolation-based methods that demand careful hyperparameter tuning, FedFew automatically discovers the optimal model diversity through its optimization process. Experiments across vision, NLP, and real-world medical imaging datasets demonstrate that FedFew, with just 3 models, consistently outperforms other state-of-the-art approaches. Code is available at https://github.com/pgg3/FedFew.

cs.AI↗

arXiv2026

Hopper-Like Growth of Higher-Order Topological Insulators

Understanding crystal growth and morphology is a fundamental issue in condensed matter physics. While crystal morphology due to the distribution and dynamics of the diffusion field has been intensively studied, how the intrinsic material properties affect crystal morphology remains unclear. In this Letter, we demonstrate that higher-order topological phases can give rise to hollowed crystal morphologies, where the corners advance faster than the central regions of the crystal, through an unconventional mechanism originating from topological electronic states. We quantitatively show this connection by analyzing both the fractal dimension $D_f$ and the fractal dimension of coastlines $D_{f,c}$. When we compare the crystals in the normal insulator and higher-order topological insulator phases with the same $D_{f}$ in the case of relatively rapid crystal growth, the former is in the dendritic shape, while the latter is in the hopper-like shape, quantified by the smaller $D_{f,c}$ in the higher-order topological phase.

cond-mat.mes-hall↗

arXiv2026

Reinforcement Learning Meets Masked Generative Models: Mask-GRPO for Text-to-Image Generation

Reinforcement learning (RL) has garnered increasing attention in text-to-image (T2I) generation. However, most existing RL approaches are tailored to either diffusion models or autoregressive models, overlooking an important alternative: masked generative models. In this work, we propose Mask-GRPO, the first method to incorporate Group Relative Policy Optimization (GRPO)-based RL into this overlooked paradigm. Our core insight is to redefine the transition probability, which is different from current approaches, and formulate the unmasking process as a multi-step decision-making problem. To further enhance our method, we explore several useful strategies, including removing the KL constraint, applying the reduction strategy, and filtering out low-quality samples. Using Mask-GRPO, we improve a base model, Show-o, with substantial improvements on standard T2I benchmarks and preference alignment, outperforming existing state-of-the-art approaches. The code is available on https://github.com/xingzhejun/Mask-GRPO

cs.CV↗

arXiv2026

Unusual Dual Flat Bands and two-dimensional Dirac-node Arc State in Kagome Metal Ni3In2S2

Kagome materials are at the frontier of condensed matter physics. An ideal kagome lattice features only one geometrically frustrated flat band spanning the entire momentum space and a single Dirac cone at the Brillouin-zone corners. However, for the first time, here we observe unusual flat-band and Dirac physics in the newly discovered "322" kagome material Ni3In2S2 by combining high-resolution synchrotron- and laser-based angle-resolved photoemission spectroscopy with a micro-focused beam, scanning tunneling microscopy, and first-principles calculations. We resolve two distinct electronic flat-band states located in close proximity to the Fermi level: a robust Topological Surface Flat Band at ~40 meV below the Fermi level on the Sulfur-terminated surface, originating from weak topological insulator states, and a kagome lattice-derived flat band at ~100 meV binding energy with an ultranarrow bandwidth (~5 meV). Instead of the single Dirac cone, the Indium-terminated surface hosts a rare two-dimensional Dirac-node arc state, where the gapless Dirac nodes extend along an open one-dimensional line crossing the Brillouin-zone boundary, exhibiting sharp linear dispersion, exceptionally high Fermi velocity, and pronounced circular dichroism. These findings establish Ni3In2S2 as a unique topological kagome metal in which multiple flat-band states of different physical origin coexist with an unusual Dirac-node arc, opening an avenue for discovering flat-band--driven and topology-enabled quantum phenomena.

cond-mat.mtrl-sci↗

arXiv2026

UACER: An Uncertainty-Adaptive Critic Ensemble Framework for Robust Adversarial Reinforcement Learning

Robust adversarial reinforcement learning has emerged as an effective paradigm for training agents to handle uncertain disturbance in real environments, with critical applications in sequential decision-making domains such as autonomous driving and robotic control. Within this paradigm, agent training is typically formulated as a zero-sum Markov game between a protagonist and an adversary to enhance policy robustness. However, the trainable nature of the adversary inevitably induces non-stationarity in the learning dynamics, leading to exacerbated training instability and convergence difficulties, particularly in high-dimensional complex environments. In this paper, we propose a novel approach, Uncertainty-Adaptive Critic Ensemble for robust adversarial Reinforcement learning (UACER), which consists of two components: 1) Diversified critic ensemble: A diverse set of K critic networks is employed in parallel to stabilize Q-value estimation in robust adversarial reinforcement learning, reducing variance and enhancing robustness compared to conventional single-critic designs. 2) Time-varying Decay Uncertainty (TDU) mechanism: Moving beyond simple linear combinations, we propose a variance-derived Q-value aggregation strategy that explicitly incorporates epistemic uncertainty to adaptively regulate the exploration-exploitation trade-off while stabilizing the training process. Comprehensive experiments across several challenging MuJoCo control problems validate the superior effectiveness of UACER, outperforming state-of-the-art methods in terms of overall performance, stability, and efficiency.

cs.LG↗

arXiv2026

AirHunt: Bridging VLM Semantics and Continuous Planning for Efficient Aerial Object Navigation

Recent advances in large Vision-Language Models (VLMs) have provided rich semantic understanding that empowers drones to search for open-set objects via natural language instructions. However, prior systems struggle to integrate VLMs into practical aerial systems due to orders-of-magnitude frequency mismatch between VLM inference and real-time planning, as well as VLMs' limited 3D scene understanding. They also lack a unified mechanism to balance semantic guidance with motion efficiency in large-scale environments. To address these challenges, we present AirHunt, an aerial object navigation system that efficiently locates open-set objects with zero-shot generalization in outdoor environments by seamlessly fusing VLM semantic reasoning with continuous path planning. AirHunt features a dual-pathway asynchronous architecture that establishes a synergistic interface between VLM reasoning and path planning, enabling continuous flight with adaptive semantic guidance that evolves through motion. Moreover, we propose an active dual-task reasoning module that exploits geometric and semantic redundancy to enable selective VLM querying, and a semantic-geometric coherent planning module that dynamically reconciles semantic priorities and motion efficiency in a unified framework, enabling seamless adaptation to environmental heterogeneity. We evaluate AirHunt across diverse object navigation tasks and environments, demonstrating a higher success rate with lower navigation error and reduced flight time compared to state-of-the-art methods. Real-world experiments further validate AirHunt's practical capability in complex and challenging environments. Code and dataset will be made publicly available before publication.

cs.RO↗

arXiv2026

Superconductivity in Electron Liquids: Precision Many-Body Treatment of Coulomb Interaction

More than a century after discovery, the theory of conventional superconductivity remains incomplete. While the importance of electron-phonon coupling is understood, a controlled first-principles treatment of Coulomb interaction is lacking. Current ab initio calculations of superconductivity rely on a phenomenological downfolding approximation, replacing Coulomb interaction with a repulsive pseudopotential μ*, and leaving ambiguities in electron-phonon coupling with dynamical Coulomb interactions unresolved. We address this via an effective field theory approach, integrating out high-energy electronic degrees of freedom using variational Diagrammatic Monte Carlo. Applied to the uniform electron gas, this establishes a microscopic procedure to implement downfolding, define the pseudopotential, and express dynamical Coulomb effects on electron-phonon coupling via the electron vertex function. We find the bare pseudopotential significantly larger than conventional values. This yields improved pseudopotential estimates in simple metals and tests density functional perturbation theory accuracy for effective electron-phonon coupling. We present an ab initio workflow computing superconducting Tc from the anomalous vertex's precursory Cooper flow. This infers Tc from normal state calculations, enabling reliable estimates of very low Tc (including near quantum phase transitions) beyond conventional reach. Validating our approach on simple metals without empirical tuning, we resolve long-standing discrepancies and predict a pressure-induced transition in Al from superconducting to non-superconducting above ~60GPa. We propose ambient-pressure Mg and Na are proximal to a similar critical point. Our work establishes a controlled ab initio framework for electron-phonon superconductivity beyond the weak-correlation limit, paving the way for reliable Tc calculations and novel material design.

cond-mat.supr-con↗

arXiv2025

Embodied Co-Design for Rapidly Evolving Agents: Taxonomy, Frontiers, and Challenges

Brain-body co-evolution enables animals to develop complex behaviors in their environments. Inspired by this biological synergy, embodied co-design (ECD) has emerged as a transformative paradigm for creating intelligent agents-from virtual creatures to physical robots-by jointly optimizing their morphologies and controllers rather than treating control in isolation. This integrated approach facilitates richer environmental interactions and robust task performance. In this survey, we provide a systematic overview of recent advances in ECD. We first formalize the concept of ECD and position it within related fields. We then introduce a hierarchical taxonomy: a lower layer that breaks down agent design into three fundamental components-controlling brain, body morphology, and task environment-and an upper layer that integrates these components into four major ECD frameworks: bi-level, single-level, generative, and open-ended. This taxonomy allows us to synthesize insights from more than one hundred recent studies. We further review notable benchmarks, datasets, and applications in both simulated and real-world scenarios. Finally, we identify significant challenges and offer insights into promising future research directions. A project associated with this survey has been created at https://github.com/Yuxing-Wang-THU/SurveyBrainBody.

cs.RO↗

arXiv2025

Agent-in-the-Loop: A Data Flywheel for Continuous Improvement in LLM-based Customer Support

We introduce an Agent-in-the-Loop (AITL) framework that implements a continuous data flywheel for iteratively improving an LLM-based customer support system. Unlike standard offline approaches that rely on batch annotations, AITL integrates four key types of annotations directly into live customer operations: (1) pairwise response preferences, (2) agent adoption and rationales, (3) knowledge relevance checks, and (4) identification of missing knowledge. These feedback signals seamlessly feed back into models' updates, reducing retraining cycles from months to weeks. Our production pilot involving US-based customer support agents demonstrated significant improvements in retrieval accuracy (+11.7% recall@75, +14.8% precision@8), generation quality (+8.4% helpfulness) and agent adoption rates (+4.5%). These results underscore the effectiveness of embedding human feedback loops directly into operational workflows to continuously refine LLM-based customer support system.

cs.AI↗

arXiv2025

Rank-Induced PL Mirror Descent: A Rank-Faithful Second-Order Algorithm for Sleeping Experts

We introduce a new algorithm, \emph{Rank-Induced Plackett--Luce Mirror Descent (RIPLM)}, which leverages the structural equivalence between the \emph{rank benchmark} and the \emph{distributional benchmark} established in \citet{BergamOzcanHsu2022}. Unlike prior approaches that operate on expert identities, RIPLM updates directly in the \emph{rank-induced Plackett--Luce (PL)} parameterization. This ensures that the algorithm's played distributions remain within the class of rank-induced distributions at every round, preserving the equivalence with the rank benchmark. To our knowledge, RIPLM is the first algorithm that is both (i) \emph{rank-faithful} and (ii) \emph{variance-adaptive} in the sleeping experts setting.

cs.LG↗

arXiv2025

Comprehensive Study of Phonon Chirality under Symmetry Constraints

Phonons are quanta of lattice vibrations, and their modes (linear, circular, or stationary) are symmetry-determined. Circularly polarized phonons, possessing nonzero angular momentum (AM), have drawn widespread attention recently. Despite widespread use of pseudo-angular momentum (PAM) and circularly polarized light polarization flips to identify chiral phonons in Raman scattering, their reliability is debated due to symmetry dependence, and experimental verification standards remain lacking. Here, we systematically study phonon chirality and associated phenomena across magnetic point groups. We establish that the AM-PAM correlation is governed by both crystalline symmetry and Wyckoff positions, dictating conditions where nonzero AM manifests in PAM signatures. Crucially, phonons belonging to distinct irreducible representations exhibit distinct experimental benchmarks, enabling direct determination of crystalline chirality and symmetry classification. Furthermore, we report the discovery of a signature for symmetry-induced phenomena, notably a half-wave plate-analogous effect induced by mirror-odd phonons. Meanwhile, we conducted five experiments to validate our theory.

cond-mat.mtrl-sci↗

arXiv2025

Fast Coherent Splitting of Bose-Einstein Condensates

Preparation of non-trivial quantum states without introducing unwanted excitations or decoherence remains a central challenge in utilizing ultracold atomic systems for quantum simulation. We employ optimal control methods to realize fast, coherent splitting of a one-dimensional Bose-Einstein condensate, achieving minimal classical excitations while preserving quantum correlations. Furthermore, we explore two-step protocols in which controlled classical motion is first induced and subsequently suppressed via tailored control sequences. Our experiments highlight the potential of optimal control for quantum state engineering and dynamical control in many-body quantum systems.

cond-mat.quant-gas↗

arXiv2025

Theory of fractional corner charges in cylindrical crystal shapes

Recent studies showed that topologically trivial insulators may have fractionally quantized corner charges due to the topological invariant called a filling anomaly. Such crystal shapes in three dimensions are restricted to vertex-transitive polyhedra, which are classified into spherical and cylindrical families. The previous works derived formulas of the fractional corner charge for the spherical family, which corresponds to the tetrahedral and cubic space groups (SGs). In this study, we derive all the corner charge formulas for the cylindrical family, which corresponds to the orthorhombic, tetragonal, hexagonal, and trigonal crystal shapes. We show that all the real-space formulas of the filling anomaly for the cylindrical SGs are universally determined by the total charges at the Wyckoff position (WP) 1a. Moreover, we derive the k-space formulas of the corner charge for the cylindrical cases with time-reversal symmetry (TRS). From our results, we also show that CsLi$_{2}$Cl_{3}, KN_{3}, and Li_{3}N are candidate materials with a quantized corner charge by using the ab initio calculations. Together with our previous work, we exhaust corner charge formulas for all the SGs and crystal shapes having quantized corner charges.

cond-mat.mtrl-sci↗

arXiv2025

Reinforcement Fine-Tuning Powers Reasoning Capability of Multimodal Large Language Models

Standing in 2025, at a critical juncture in the pursuit of Artificial General Intelligence (AGI), reinforcement fine-tuning (RFT) has demonstrated significant potential in enhancing the reasoning capability of large language models (LLMs) and has led to the development of cutting-edge AI models such as OpenAI-o1 and DeepSeek-R1. Moreover, the efficient application of RFT to enhance the reasoning capability of multimodal large language models (MLLMs) has attracted widespread attention from the community. In this position paper, we argue that reinforcement fine-tuning powers the reasoning capability of multimodal large language models. To begin with, we provide a detailed introduction to the fundamental background knowledge that researchers interested in this field should be familiar with. Furthermore, we meticulously summarize the improvements of RFT in powering reasoning capability of MLLMs into five key points: diverse modalities, diverse tasks and domains, better training algorithms, abundant benchmarks and thriving engineering frameworks. Finally, we propose five promising directions for future research that the community might consider. We hope that this position paper will provide valuable insights to the community at this pivotal stage in the advancement toward AGI. Summary of works done on RFT for MLLMs is available at https://github.com/Sun-Haoyuan23/Awesome-RL-based-Reasoning-MLLMs.

cs.CL↗