arXiv subjects

Zhenzhen Li

Publications and source records attributed to Zhenzhen Li.

At least 19 recordsLinked to original sources

arXiv2026

Trusting Imagined Futures: Auditing Imagined Rewards for VLA Post-Training

Video world models enable candidate actions to be evaluated through imagined outcomes, but high predicted rewards can favor trajectories with poor execution quality. We introduce Trusting Imagined Futures (TIF), an offline supervision selection method for vision-language-action (VLA) post-training that audits predicted rewards using execution feedback. A multidimensional reward model predicts a utility score for candidate ranking and five dimensions of execution quality. An auditor uses reward predictions and uncertainty to identify overestimation patterns learned from paired imagined videos and real executions. TIF accepts the candidate with the highest predicted utility when its advantage over the existing trajectory exceeds a margin and its audit risk passes a threshold; otherwise, it retains the existing supervision. The retained real trajectories then supervise behavior cloning. Across four LIBERO suites and three training seeds, TIF achieves a mean success rate of 76.25%, improving over training on the existing trajectories by 3.00 percentage points with the same trajectory count and training budget. Adding audit gating yields 1.47 percentage points with the reward model and advantage margin held fixed. The gains persist against advantage ranking and logistic regression risk selection at matched replacement counts. On recorded videos from three real robot tasks, TIF increases clean completion among selected trajectories from 58.33% to 70.00% across 60 candidate pools.

cs.RO↗

arXiv2026

Vector-Quantized World Model: Latent Feedback Motion Planning for Robot Manipulation

Visual world models have shown strong potential for robot planning, yet long-horizon manipulation remains challenging due to compounding prediction errors, large search complexity, and latent plans that the low-level controller cannot reliably reach. In this work, we introduce Vector-Quantized World Model (VQ-WM), a latent feedback motion planning framework that integrates discrete long-horizon search with continuous visual feedback control: a discrete latent predictor models transitions between vector-quantized latent states, allowing the planner to look far ahead and find a shortest path to the goal through discrete search and a continuous learned controller sequentially reaches the planned states through visual feedback, correcting deviations to mitigate compounding errors. Whereas most vector quantization uses Euclidean distance, we use a novel estimator of the first-passage-time as a quasi-metric, which aligns the planning abstraction with the controller's capabilities. We evaluate on the physical robot manipulation of the Tower of Hanoi puzzle: whereas existing state-of-the-art foundation models (VLAs, WAMs) only succeed when the solution trace in their training dataset, our method achieves 100\% success rate with only short-play data, demonstrate its ability to compose long-horizon solutions beyond the short demonstrated trajectories and execute them reliably. Our solution also outperforms strongest world-models by 13.3--85.3 percentage points on simulation benchmarks spanning manipulation and locomotion, demonstrating consistent progress and new capabilities.

cs.RO↗

arXiv2026

FLEX-WAM: Flexible Block-Causal World-Action Models for Long-Horizon Imagination and Planning

World--action models (WAMs) promise a unified model that predicts action-conditioned futures, generates feasible actions, and supports planning in imagination. However, existing joint video--action models often use computationally heavy, fixed-horizon backbones ill-suited to streaming inference and stable long-horizon open-loop rollouts. We introduce FLEX-WAM, a Flexible and Efficient Block-Causal World--Action Model for unified simulation and policy inference. FLEX-WAM supports variable-length contexts and non-causal prediction horizons, as well as infinite autoregressive generation frame by frame or block by block. Its block-causal, KV-cacheable architecture combines axial attention and blockwise diffusion forcing to enable efficient real-time rollout and deployment-time latency--throughput tradeoffs without retraining. Joint training can nevertheless produce plausible futures that weakly respond to commanded actions. We address this failure mode by balancing state and action flow-matching gradient contributions across the state--action diffusion-noise grid and regulating world-model sampling using Forward-Dynamics (FD) elasticity, an efficient training-time proxy for action responsiveness. Across simulated and real-world datasets, FLEX-WAM achieves superior multi-step prediction quality and latency while producing stable joint state--action rollouts for thousands of steps. As a joint action proposer and simulator within MCTS, it solves long-horizon PushT and all five OGBench Puzzle-4x4 tasks entirely in imagination. On a bimanual OpenArm-based robot, a single checkpoint jointly serves as a play policy and expected-outcome predictor, enabling real-time identification and collection of model--reality mismatches for future self-improvement.

cs.RO↗

arXiv2026

Diffusion contrast induced dynamics in a time-periodic competition-diffusion system with equal total resources

This paper investigates a time-periodic, spatiotemporally heterogeneous competition-diffusion system, where two species have different intrinsic growth rates but equal total resources per period. Using asymptotic analysis and principal eigenvalue theory, we systematically examine its dynamics. When both diffusion rates are small, spatial heterogeneity of the difference in time-averaged resources leads to uniform persistence, stable coexistence, and asymptotic spatial segregation; whereas its spatial uniformity produces multiple regimes--fast-diffuser selection, parameter-dependent coexistence, or slow-diffuser dominance--separated by smooth threshold curves under different structural conditions. In the regime where at least one diffusion rate is large, we obtain a detailed classification of the local dynamics, with sharp delineation of parameter regions for stable coexistence, competitive exclusion, and a novel bistable scenario, where both semi-trivial periodic solutions are linearly stable while an unstable coexistence state also exists. We further characterize the asymptotic profiles of positive solutions in the mixed-scale diffusion limit. Our results show that the interplay between temporal periodicity and spatial heterogeneity can overturn the classical "slower diffuser always prevails" principle, leading to a substantially richer range of ecological outcomes.

math.AP↗

arXiv2026

Reason Before You Retrieve: Agentic Planning for Multi-modal RAG

Multimodal retrieval-augmented generation (mRAG) aims to answer image-text queries with external knowledge, but most existing systems still retrieve directly from raw multimodal input over a flat evidence space. This design often struggles with two key challenges: the retrieval target is under-specified because the question intent must be grounded to the correct visual referent, and the search space is weakly structured, forcing semantically distinct evidence to compete in a single global ranking step. We propose MM-R2, a multimodal agentic retrieval framework that reasons before retrieval by explicitly modeling both what to retrieve and where to search. MM-R2 first constructs an intent-grounded retrieval state from the image-question pair, capturing the information need, grounded referent, and retrieval constraints. It then performs retrieval over a structured KnowledgeMap, where the agent selects relevant retrieval units before issuing grounded queries within them. To enable this capability, we build MM-R2-Traj, a large-scale trajectory dataset of multi-step retrieval processes, and adopt a two-stage post-training strategy with supervised fine-tuning and GRPO. Experiments on Infoseek and Encyclopedic VQA datasets show that MM-R2 substantially outperforms strong baselines on answer accuracy while also yielding more interpretable and verifiable retrieval trajectories.

cs.AI↗

arXiv2026

The Narrow-Line Seyfert 1 Phenomenon: Accretion State Versus Host Galaxy Properties

The physical origin of the narrow-line Seyfert 1 (NLS1) and broad-line Seyfert 1 (BLS1) dichotomy remains debated, with competing scenarios invoking host-galaxy evolution or intrinsic accretion physics. We analysed host-galaxy properties and AGN luminosities obtained from CIGALE spectral energy distribution fitting for $\sim$12,000 Type 1 AGNs from the Sloan Digital Sky Survey, of which 29\% are NLS1s. Globally, NLS1s have lower virial black hole masses, higher inferred Eddington ratios, lower stellar masses, and higher specific star formation rates than BLS1s. In the FWHM(\Hb)--$L_{\rm AGN}$ plane, the conventional 2000 km s$^{-1}$ boundary is better viewed as an empirical division within a continuous parameter space rather than a physical threshold, with Fe~II tracing the high-accretion end. In a host-matched subsample of 767 NLS1--BLS1 pairs with statistically indistinguishable stellar mass, black hole mass, and redshift, NLS1s still show higher Eddington ratios, stronger Fe~II emission, and bluer optical continua, together with elevated SFR and dust attenuation, suggesting that the NLS1 phenomenon is most naturally associated with a high-accretion state within the continuous distribution of Type 1 AGNs, while host-galaxy gas supply may also play a role in modulating its strength. In this picture, NLS1 and BLS1 classifications reflect different locations within a continuous accretion sequence of the same underlying population rather than two physically disjoint classes.

astro-ph.GA↗

arXiv2025

Time-series Forecast for Indoor Zone Air Temperature with Long Horizons: A Case Study with Sensor-based Data from a Smart Building

With the press of global climate change, extreme weather and sudden weather changes are becoming increasingly common. To maintain a comfortable indoor environment and minimize the contribution of the building to climate change as much as possible, higher requirements are placed on the operation and control of HVAC systems, e.g., more energy-efficient and flexible to response to the rapid change of weather. This places demands on the rapid modeling and prediction of zone air temperatures of buildings. Compared to the traditional simulation-based approach such as EnergyPlus and DOE2, a hybrid approach combined physics and data-driven is more suitable. Recently, the availability of high-quality datasets and algorithmic breakthroughs have driven a considerable amount of work in this field. However, in the niche of short- and long-term predictions, there are still some gaps in existing research. This paper aims to develop a time series forecast model to predict the zone air temperature in a building located in America on a 2-week horizon. The findings could be further improved to support intelligent control and operation of HVAC systems (i.e. demand flexibility) and could also be used as hybrid building energy modeling.

cs.LG↗

arXiv2025

Catch Me If You Can? Not Yet: LLMs Still Struggle to Imitate the Implicit Writing Styles of Everyday Authors

As large language models (LLMs) become increasingly integrated into personal writing tools, a critical question arises: can LLMs faithfully imitate an individual's writing style from just a few examples? Personal style is often subtle and implicit, making it difficult to specify through prompts yet essential for user-aligned generation. This work presents a comprehensive evaluation of state-of-the-art LLMs' ability to mimic personal writing styles via in-context learning from a small number of user-authored samples. We introduce an ensemble of complementary metrics-including authorship attribution, authorship verification, style matching, and AI detection-to robustly assess style imitation. Our evaluation spans over 40000 generations per model across domains such as news, email, forums, and blogs, covering writing samples from more than 400 real-world authors. Results show that while LLMs can approximate user styles in structured formats like news and email, they struggle with nuanced, informal writing in blogs and forums. Further analysis on various prompting strategies such as number of demonstrations reveal key limitations in effective personalization. Our findings highlight a fundamental gap in personalized LLM adaptation and the need for improved techniques to support implicit, style-consistent generation. To aid future research and for reproducibility, we open-source our data and code.

cs.CL↗

arXiv2025

Extended Emission-line Region in a Poststarburst Galaxy Hosting Tidal Disruption Event AT2019qiz and Quasiperiodic Eruptions

We present a comprehensive analysis of the extended emission line region (EELR) in the host galaxy of the tidal disruption event (TDE) AT2019qiz, utilizing VLT/MUSE integral-field spectroscopy. The high spatial-resolution data reveal a bi-conical emission structure approximately $3.7~\mathrm{kpc}$ in scale within the galactic center, characterized by a prominent [OIII] line in the nucleus and significant [NII] line emission extending into the EELR. Spectral analysis of the EELR indicates line ratios consistent with Seyfert ionization in the center and LINER-type ionization in the outer diffuse region, suggesting ionization from galactic nuclear activity. The required ionizing luminosity, estimated from the H$\rmα$ and H$\rmβ$ luminosities based on the photoionization and recombination balance assumption, is $10^{41.8}$ $\mathrm{erg\,s^{-1}}$ for all spaxels classified as active galactic nucleus (AGN), and $10^{40.7}$ $\mathrm{erg\,s^{-1}}$ for spaxels in the central $0.9~\mathrm{kpc}$ Seyfert region. However, the current bolometric luminosity of the nucleus $L_{\text{bol}} \leq 10^{40.8}\,\mathrm{erg\,s^{-1}}$, estimated from quiescent-state soft X-ray observations, is insufficient to ionize the entire EELR, implying a recently faded AGN or a delayed response to historical activity. Stellar population analysis reveals a post-starburst characteristic in the EELR, and the gas kinematics show disturbances and non-circular components compared to the stellar kinematics. Notably, the recent detection of quasi-periodic eruptions (QPEs) in the X-ray light curve of AT2019qiz confirms the TDE-QPE association. Our findings provide direct evidence for an AGN-like EELR in the host galaxy of the nearest TDE with QPE detection, offering new insights into the complex interplay between TDEs, QPEs, AGN activity, and host galaxy evolution.

astro-ph.GA↗

arXiv2025

Human2LocoMan: Learning Versatile Quadrupedal Manipulation with Human Pretraining

Quadrupedal robots have demonstrated impressive locomotion capabilities in complex environments, but equipping them with autonomous versatile manipulation skills in a scalable way remains a significant challenge. In this work, we introduce a cross-embodiment imitation learning system for quadrupedal manipulation, leveraging data collected from both humans and LocoMan, a quadruped equipped with multiple manipulation modes. Specifically, we develop a teleoperation and data collection pipeline, which unifies and modularizes the observation and action spaces of the human and the robot. To effectively leverage the collected data, we propose an efficient modularized architecture that supports co-training and pretraining on structured modality-aligned data across different embodiments. Additionally, we construct the first manipulation dataset for the LocoMan robot, covering various household tasks in both unimanual and bimanual modes, supplemented by a corresponding human dataset. We validate our system on six real-world manipulation tasks, where it achieves an average success rate improvement of 41.9% overall and 79.7% under out-of-distribution (OOD) settings compared to the baseline. Pretraining with human data contributes a 38.6% success rate improvement overall and 82.7% under OOD settings, enabling consistently better performance with only half the amount of robot data. Our code, hardware, and data are open-sourced at: https://human2bots.github.io.

cs.RO↗

arXiv2025

FedRPCA: Enhancing Federated LoRA Aggregation Using Robust PCA

LoRA has emerged as one of the most promising fine-tuning techniques, especially for federated learning (FL), since it significantly reduces communication and computation costs at resource-constrained clients. However, data heterogeneity remains a significant challenge for LoRA-based FL, and the conventional aggregation strategy based on FedAvg suffers from slow convergence and suboptimal accuracy. Motivated by recent advances in model merging, particularly Task Arithmetic, we explore the idea of aggregating client LoRA parameters using scaled averaging. We first observe that a naive application of Task Arithmetic is ineffective due to the high cosine similarity between client updates, indicating significant common knowledge in the updates across clients. To address this issue, we propose decomposing client LoRA updates via Robust Principal Component Analysis (Robust-PCA) into a common low-rank component and client-specific sparse components. Our proposed algorithm FedRPCA aggregates the low-rank components through averaging, consolidating common knowledge, and applies scaled averaging to the sparse components to amplify client-specific knowledge. We evaluate our approach across a variety of vision and language tasks and demonstrate that it achieves higher final accuracy and faster convergence compared to competing baselines.

cs.LG↗

arXiv2024

Serendipitous Catch of a Giant Jellyfish: an Ionized Nebula around 3C 275.1 with 170 kpc Long Tails

3C 275.1 is a blue quasar at $z=0.55522$, hosting powerful outflows and residing in a complex environment. We present a serendipitously detected giant nebula surrounding 3C 275.1, which shows morphological features resembling those of objects known as "jellyfish galaxy", with extremely long tails of ionized gas extending to 170 kpc in projection. We analyze its optical spectra taken by the MUSE on the VLT. The brighter part of this giant nebula exceeds 100 kpc, whose rotation curve does not flatten out, is very different from those of normal spiral galaxies. This system shares some characteristics common to those formed via ram pressure stripping (RPS), yet its long narrow tails and higher ionization are unusual compared to known tails in jellyfish galaxies, not fully consistent with a simple RPS scenario. Our photoionization simulation and the inferred short recombination timescale both suggest that besides the quasar 3C 275.1, an extra source of ionization is necessary to keep the gas ionized at such distance from the nucleus, which could be related to RPS, tidal interaction or AGN outflow, providing new evidence of active dynamical interaction of a galaxy with the intracluster medium.

astro-ph.GA↗

arXiv2024

Auto-Train-Once: Controller Network Guided Automatic Network Pruning from Scratch

Current techniques for deep neural network (DNN) pruning often involve intricate multi-step processes that require domain-specific expertise, making their widespread adoption challenging. To address the limitation, the Only-Train-Once (OTO) and OTOv2 are proposed to eliminate the need for additional fine-tuning steps by directly training and compressing a general DNN from scratch. Nevertheless, the static design of optimizers (in OTO) can lead to convergence issues of local optima. In this paper, we proposed the Auto-Train-Once (ATO), an innovative network pruning algorithm designed to automatically reduce the computational and storage costs of DNNs. During the model training phase, our approach not only trains the target model but also leverages a controller network as an architecture generator to guide the learning of target model weights. Furthermore, we developed a novel stochastic gradient algorithm that enhances the coordination between model training and controller network training, thereby improving pruning performance. We provide a comprehensive convergence analysis as well as extensive experiments, and the results show that our approach achieves state-of-the-art performance across various model architectures (including ResNet18, ResNet34, ResNet50, ResNet56, and MobileNetv2) on standard benchmark datasets (CIFAR-10, CIFAR-100, and ImageNet).

cs.CV↗

arXiv2023

Leveraging Foundation Models to Improve Lightweight Clients in Federated Learning

Federated Learning (FL) is a distributed training paradigm that enables clients scattered across the world to cooperatively learn a global model without divulging confidential data. However, FL faces a significant challenge in the form of heterogeneous data distributions among clients, which leads to a reduction in performance and robustness. A recent approach to mitigating the impact of heterogeneous data distributions is through the use of foundation models, which offer better performance at the cost of larger computational overheads and slower inference speeds. We introduce foundation model distillation to assist in the federated training of lightweight client models and increase their performance under heterogeneous data settings while keeping inference costs low. Our results show improvement in the global model performance on a balanced testing set, which contains rarely observed samples, even under extreme non-IID client data distributions. We conduct a thorough evaluation of our framework with different foundation model backbones on CIFAR10, with varying degrees of heterogeneous data distributions ranging from class-specific data partitions across clients to dirichlet data sampling, parameterized by values between 0.01 and 1.0.

cs.LG↗

arXiv2023

Text-driven Prompt Generation for Vision-Language Models in Federated Learning

Prompt learning for vision-language models, e.g., CoOp, has shown great success in adapting CLIP to different downstream tasks, making it a promising solution for federated learning due to computational reasons. Existing prompt learning techniques replace hand-crafted text prompts with learned vectors that offer improvements on seen classes, but struggle to generalize to unseen classes. Our work addresses this challenge by proposing Federated Text-driven Prompt Generation (FedTPG), which learns a unified prompt generation network across multiple remote clients in a scalable manner. The prompt generation network is conditioned on task-related text input, thus is context-aware, making it suitable to generalize for both seen and unseen classes. Our comprehensive empirical evaluations on nine diverse image classification datasets show that our method is superior to existing federated prompt learning methods, that achieve overall better generalization on both seen and unseen classes and is also generalizable to unseen datasets.

cs.CV↗

arXiv2023

Cross-corpus Readability Compatibility Assessment for English Texts

Text readability assessment has gained significant attention from researchers in various domains. However, the lack of exploration into corpus compatibility poses a challenge as different research groups utilize different corpora. In this study, we propose a novel evaluation framework, Cross-corpus text Readability Compatibility Assessment (CRCA), to address this issue. The framework encompasses three key components: (1) Corpus: CEFR, CLEC, CLOTH, NES, OSP, and RACE. Linguistic features, GloVe word vector representations, and their fusion features were extracted. (2) Classification models: Machine learning methods (XGBoost, SVM) and deep learning methods (BiLSTM, Attention-BiLSTM) were employed. (3) Compatibility metrics: RJSD, RRNSS, and NDCG metrics. Our findings revealed: (1) Validated corpus compatibility, with OSP standing out as significantly different from other datasets. (2) An adaptation effect among corpora, feature representations, and classification methods. (3) Consistent outcomes across the three metrics, validating the robustness of the compatibility assessment framework. The outcomes of this study offer valuable insights into corpus selection, feature representation, and classification methods, and it can also serve as a beginning effort for cross-corpus transfer learning.

cs.CL↗

arXiv2022

Long-time behavior of a nonlocal dispersal logistic model with seasonal succession

This paper is devoted to a nonlocal dispersal logistic model with seasonal succession in one-dimensional bounded habitat, where the seasonal succession accounts for the effect of two different seasons. Firstly, we provide the persistence-extinction criterion for the species, which is different from that for local diffusion model. Then we show the asymptotic profile of the time-periodic positive solution as the species persists in long run.

math.AP↗

arXiv2022

Asymptotic Escape of Spurious Critical Points on the Low-rank Matrix Manifold

We show that on the manifold of fixed-rank and symmetric positive semi-definite matrices, the Riemannian gradient descent algorithm almost surely escapes some spurious critical points on the boundary of the manifold. Our result is the first to partially overcome the incompleteness of the low-rank matrix manifold without changing the vanilla Riemannian gradient descent algorithm. The spurious critical points are some rank-deficient matrices that capture only part of the eigen components of the ground truth. Unlike classical strict saddle points, they exhibit very singular behavior. We show that using the dynamical low-rank approximation and a rescaled gradient flow, some of the spurious critical points can be converted to classical strict saddle points in the parameterized domain, which leads to the desired result. Numerical experiments are provided to support our theoretical findings.

math.OC↗