Yizhou Zhang

Yizhou Zhang, Defu Cao, Yan Liu

Recent years have witnessed the rise of misinformation campaigns that spread specific narratives on social media to manipulate public opinions on different areas, such as politics and healthcare. Consequently, an effective and efficient automatic methodology to estimate the influence of the misinformation on user beliefs and activities is needed. However, existing works on misinformation impact estimation either rely on small-scale psychological experiments or can only discover the correlation between user behaviour and misinformation. To address these issues, in this paper, we build up a causal framework that model the causal effect of misinformation from the perspective of temporal point process. To adapt the large-scale data, we design an efficient yet precise way to estimate the Individual Treatment Effect(ITE) via neural temporal point process and gaussian mixture models. Extensive experiments on synthetic dataset verify the effectiveness and efficiency of our model. We further apply our model on a real-world dataset of social media posts and engagements about COVID-19 vaccines. The experimental results indicate that our model recognized identifiable causal effect of misinformation that hurts people's subjective emotions toward the vaccines.

Yizhou Zhang, Loc Trinh, Defu Cao et al.

Recent years have witnessed the sustained evolution of misinformation that aims at manipulating public opinions. Unlike traditional rumors or fake news editors who mainly rely on generated and/or counterfeited images, text and videos, current misinformation creators now more tend to use out-of-context multimedia contents (e.g. mismatched images and captions) to deceive the public and fake news detection systems. This new type of misinformation increases the difficulty of not only detection but also clarification, because every individual modality is close enough to true information. To address this challenge, in this paper we explore how to achieve interpretable cross-modal de-contextualization detection that simultaneously identifies the mismatched pairs and the cross-modal contradictions, which is helpful for fact-check websites to document clarifications. The proposed model first symbolically disassembles the text-modality information to a set of fact queries based on the Abstract Meaning Representation of the caption and then forwards the query-image pairs into a pre-trained large vision-language model select the ``evidences" that are helpful for us to detect misinformation. Extensive experiments indicate that the proposed methodology can provide us with much more interpretable predictions while maintaining the accuracy same as the state-of-the-art model on this task.

Yizhou Zhang, Guannan Qu, Pan Xu et al.

We study a multi-agent reinforcement learning (MARL) problem where the agents interact over a given network. The goal of the agents is to cooperatively maximize the average of their entropy-regularized long-term rewards. To overcome the curse of dimensionality and to reduce communication, we propose a Localized Policy Iteration (LPI) algorithm that provably learns a near-globally-optimal policy using only local information. In particular, we show that, despite restricting each agent's attention to only its $κ$-hop neighborhood, the agents are able to learn a policy with an optimality gap that decays polynomially in $κ$. In addition, we show the finite-sample convergence of LPI to the global optimal policy, which explicitly captures the trade-off between optimality and computational complexity in choosing $κ$. Numerical simulations demonstrate the effectiveness of LPI.

Defu Cao, Wen Ye, Yizhou Zhang et al.

Foundation models, particularly Large Language Models (LLMs), have revolutionized text and video processing, yet time series data presents distinct challenges for such approaches due to domain-specific features such as missing values, multi-resolution characteristics, etc. Furthermore, the de-facto autoregressive transformers tend to learn deterministic temporal dependencies within pre-trained data while overlooking inherent uncertainties and lacking integration of physical constraints. In this paper, we introduce TimeDiT, a diffusion transformer model that synergistically combines transformer-based temporal dependency learning with diffusion-based probabilistic sampling. TimeDiT employs a unified masking mechanism to harmonize the training and inference process across diverse tasks while introducing a theoretically grounded, finetuning-free model editing strategy that enables flexible integration of external knowledge during sampling. Acknowledging the challenges of unifying multiple downstream tasks under a single model, our systematic evaluation demonstrates TimeDiT's effectiveness both in fundamental tasks, i.e., forecasting and imputation, through zero-shot/fine-tuning; and in domain tasks, i.e., multi-resolution forecasting, anomaly detection, and data generation, establishing it as a \textit{proto-foundation model} that bridges the gap between general-purpose and domain-specific models.

Wen Ye, Wei Yang, Defu Cao et al.

Time series analysis is crucial in real-world applications, yet traditional methods focus on isolated tasks only, and recent studies on time series reasoning remain limited to either single-step inference or are constrained to natural language answers. In this work, we introduce TS-Reasoner, a domain-specialized agent designed for multi-step time series inference. By integrating large language model (LLM) reasoning with domain-specific computational tools and an error feedback loop, TS-Reasoner enables domain-informed, constraint-aware analytical workflows that combine symbolic reasoning with precise numerical analysis. We assess the system's capabilities along two axes: (1) fundamental time series understanding assessed by TimeSeriesExam and (2) complex, multi-step inference evaluated by a newly proposed dataset designed to test both compositional reasoning and computational precision in time series analysis. Experiments show that our approach outperforms standalone general-purpose LLMs in both basic time series concept understanding as well as the multi-step time series inference task, highlighting the promise of domain-specialized agents for automating real-world time series reasoning and analysis.

Chenyou Guo, Zongqi Liu, Yizhou Zhang et al.

While general-purpose Large Language Models (LLMs) applied to Geology often hallucinate when reasoning about subsurface structures and deep-time evolution, current AI in Earth sciences predominantly targets surface remote sensing and GIS. To bridge this gap, we introduce Geo-Expert, a family of parameter-efficient geological LLMs fine-tuned on a custom-curated, high-quality instruction dataset processed using our custom instruction synthesis pipeline. We investigate the impact of model scaling and architecture by fine-tuning three base models: Qwen3-8B, Qwen3-32B, and Gemma-3-27B, with Low-Rank Adaptation (LoRA) method. Our extensive evaluation on a novel domain-specific benchmark, Geo-Eval, reveals that a domain-aligned 8B model can outperform open-weight 70B generalists and proprietary GPT-4o on specialized geological reasoning, while a 32B variant approaches frontier reasoning models. The optimized 8B model further offers a competitive cost-performance ratio for deployment. This work provides a reproducible recipe for democratizing scientific LLMs and establishes a baseline for geological artificial intelligence.

Yizhou Zhang, Ning Lv, Teng Wang et al.

Group relative policy optimization (GRPO) has demonstrated significant potential in improving the reasoning capabilities of large language models (LLMs) via reinforcement learning. However, its practical deployment is impeded by an excessively slow training process, primarily attributed to the computationally intensive autoregressive generation of multiple responses per query, which makes the generation phase the primary performance bottleneck. Although speculative decoding presents a promising direction for acceleration, its direct application in GRPO achieves limited speedup under high-concurrency training conditions. To overcome this limitation, we propose a concurrency-aware speculative decoding framework that dynamically adjusts the drafting and verification strategy according to real-time concurrency levels, thereby maximizing the acceleration of the generation process. Furthermore, to address performance degradation arising from distributional drift between the evolving target model and the fixed draft model during training, we introduce an online draft learning mechanism that enables the draft model to continuously adapt using feedback signals from the target model. Experimental results across multiple mathematical reasoning datasets and models demonstrate that the proposed method achieves end-to-end speedups of 2.35x to 2.72x, significantly surpassing baseline approaches in efficiency. The code is available at https://github.com/yedaotian9/GRPO_speculative.

Jisheng Dang, Yizhou Zhang, Hao Ye et al.

Fine-grained video captioning aims to generate detailed, temporally coherent descriptions of video content. However, existing methods struggle to capture subtle video dynamics and rich detailed information. In this paper, we leverage preference learning to enhance the performance of vision-language models in fine-grained video captioning, while mitigating several limitations inherent to direct preference optimization (DPO). First, we propose a pipeline for constructing preference pairs that leverages the intrinsic properties of VLMs along with partial assistance from large language models, achieving an optimal balance between cost and data quality. Second, we propose Synergistic Preference Optimization (SynPO), a novel optimization method offering significant advantages over DPO and its variants. SynPO prevents negative preferences from dominating the optimization, explicitly preserves the model's language capability to avoid deviation of the optimization objective, and improves training efficiency by eliminating the need for the reference model. We extensively evaluate SynPO not only on video captioning benchmarks (e.g., VDC, VDD, VATEX) but also across well-established NLP tasks, including general language understanding and preference evaluation, using diverse pretrained models. Results demonstrate that SynPO consistently outperforms DPO variants while achieving 20\% improvement in training efficiency. Code is available at https://github.com/longmalongma/SynPO

Hao Ye, Jisheng Dang, Junfeng Fang et al.

Recent extensive research has demonstrated that the enhanced reasoning capabilities acquired by models through Reinforcement Learning with Verifiable Rewards (RLVR) are primarily concentrated within the rank-1 components. Predicated on this observation, we employed Periodic Rank-1 Substitution and identified a counterintuitive phenomenon: RLVR may exhibit implicit reward overfitting to the training dataset. Specifically, the model can achieve satisfactory performance on the test set even when its rewards remain relatively low during the training process. Furthermore, we characterize three distinct properties of RL training: (1) The effective rank-1 component in RLVR don't maintain other model knowledge except mathematical reasoning capability. (2) RLVR fundamentally functions by optimizing a specific singular spectrum. The distribution of singular values of almost all linear layers in RLVR-trained model behaves like heavy-tailed distribution. (3) the left singular vectors associated with rank-1 components demonstrate a stronger alignment tendency during training, which echoes the discovery that RLVR is optimizing sampling efficiency in essence. Taken together, our findings and analysis further reveal how RLVR shapes model parameters and offer potential insights for improving existing RL paradigms or other training paradigms to implement continual learning.

Yizhou Zhang, Eric Mazumdar

Learning stationary policies in infinite-horizon general-sum Markov games (MGs) remains a fundamental open problem in Multi-Agent Reinforcement Learning (MARL). While stationary strategies are preferred for their practicality, computing stationary forms of classic game-theoretic equilibria is computationally intractable -- a stark contrast to the comparative ease of solving single-agent RL or zero-sum games. To bridge this gap, we study Risk-averse Quantal response Equilibria (RQE), a solution concept rooted in behavioral game theory that incorporates risk aversion and bounded rationality. We demonstrate that RQE possesses strong regularity conditions that make it uniquely amenable to learning in MGs. We propose a novel two-timescale Actor-Critic algorithm characterized by a fast-timescale actor and a slow-timescale critic. Leveraging the regularity of RQE, we prove that this approach achieves global convergence with finite-sample guarantees. We empirically validate our algorithm in several environments to demonstrate superior convergence properties compared to risk-neutral baselines.

Zhan Jin, Yu Luo, Yizhou Zhang et al.

Conventional pixel-wise loss functions fail to enforce topological constraints in coronary vessel segmentation, producing fragmented vascular trees despite high pixel-level accuracy. We present ARIADNE, a two-stage framework coupling preference-aligned perception with RL-based diagnostic reasoning for topologically coherent stenosis detection. The perception module employs DPO to fine-tune the Sa2VA vision-language foundation model using Betti number constraints as preference signals, aligning the policy toward geometrically complete vessel structures rather than pixel-wise overlap metrics. The reasoning module formulates stenosis localization as a Markov Decision Process with an explicit rejection mechanism that autonomously defers ambiguous anatomical candidates such as bifurcations and vessel crossings, shifting from coverage maximization to reliability optimization. On 1,400 clinical angiograms, ARIADNE achieves state-of-the-art centerline Dice of 0.838, reduces false positives by 41% compared to geometric baselines. External validation on multi-center benchmarks ARCADE and XCAD confirms generalization across acquisition protocols. This represents the first application of DPO for topological alignment in medical imaging, demonstrating that preference-based learning over structural constraints mitigates topological violations while maintaining diagnostic sensitivity in interventional cardiology workflows.

Yizhou Zhang, Lun Du, Defu Cao et al.

Foundation models, such as Large language Models (LLMs), have attracted significant amount of interest due to their large number of applications. However, when handling tasks involving repetitive sub-tasks and/or deceptive contents, such as arithmetic calculation and article-level fake news detection, simple instructional prompts suffer from inaccurate responses. Existing works show that more complicated prompting strategies, such as Chain-of-Thoughts and Least-to-Most, can unlock LLM's powerful capacity in diverse areas. Recent researches reveal that simple divide-and-conquer prompting strategy, i.e. simply dividing the input sequence to multiple sub-inputs, can also substantially improve LLM's performance in some specific tasks such as misinformation detection. In this paper, we aim at examining the utility of divide-and-conquer prompting strategy and answer on which kind of tasks this strategy gets advantages. Specifically, we provide a theoretic analysis to divide-and-conquer prompting strategy and help us identify the specific tasks where DaC prompting can bring performance boost with theoretic guarantee. We then present two cases (large integer arithmetic and fact verification) where experimental results aligns with our theoretic analysis.

Chengrui Qu, Yizhou Zhang, Nicholas Lanzetti et al.

Many emerging agentic paradigms require agents to collaborate with one another (or people) to achieve shared goals. Unfortunately, existing approaches to learning policies for such collaborative problems produce brittle solutions that fail when paired with new partners. We attribute these failures to a combination of free-riding during training and a lack of strategic robustness. To address these problems, we study the concept of strategic risk aversion and interpret it as a principled inductive bias for generalizable cooperation with unseen partners. While strategically risk-averse players are robust to deviations in their partner's behavior by design, we show that, in collaborative games, they also (1) can have better equilibrium outcomes than those at classical game-theoretic concepts like Nash, and (2) exhibit less or no free-riding. Inspired by these insights, we develop a multi-agent reinforcement learning (MARL) algorithm that integrates strategic risk aversion into standard policy optimization methods. Our empirical results across collaborative benchmarks (including an LLM collaboration task) validate our theory and demonstrate that our approach consistently achieves reliable collaboration with heterogeneous and previously unseen partners across collaborative tasks.

Yizhou Zhang

Empirical scaling laws describe how test loss and other performance metrics depend on model size, dataset size, and compute. While such laws are consistent within specific regimes, apparently distinct scaling behaviors have been reported for related settings such as model compression. Motivated by recent progress in spectral analyses of neural representations, this paper develops a \emph{generalized spectral framework} that unifies learning dynamics and compression phenomena under a common functional ansatz. We generalize the spectral evolution function from the linear kernel form $g(λt)=λt$ to an asymptotically polynomial function $g(λ,t;β)$, characterized by an effective spectral--temporal elasticity $ρ(β)$. This framework recovers existing lazy and feature-learning theories as special cases and yields an invariant relation between learning and compression

Yizhou Zhang, Kishan Panaganti, Laixi Shi et al.

Differential Privacy (DP) provides a rigorous framework for privacy, ensuring the outputs of data-driven algorithms remain statistically indistinguishable across datasets that differ in a single entry. While guaranteeing DP generally requires explicitly injecting noise either to the algorithm itself or to its outputs, the intrinsic randomness of existing algorithms presents an opportunity to achieve DP ``for free''. In this work, we explore the role of regularization in achieving DP across three different decision-making problems: multi-armed bandits, linear contextual bandits, and reinforcement learning from human feedback (RLHF), in offline data settings. We show that adding KL-regularization to the learning objective (a common approach in optimization algorithms) makes the action sampled from the resulting stochastic policy itself differentially private. This offers a new route to privacy guarantees without additional noise injection, while also preserving the inherent advantage of regularization in enhancing performance.

Yizhou Zhang

Neural scaling laws and double-descent phenomena suggest that deep-network training obeys a simple macroscopic structure despite highly nonlinear optimization dynamics. We derive such structure directly from gradient descent in function space. For mean-squared error loss, the training error evolves as $\dot e_t=-M(t)e_t$ with $M(t)=J_{θ(t)}J_{θ(t)}^{\!*}$, a time-dependent self-adjoint operator induced by the network Jacobian. Using Kato perturbation theory, we obtain an exact system of coupled modewise ODEs in the instantaneous eigenbasis of $M(t)$. To extract macroscopic behavior, we introduce a logarithmic spectral-shell coarse-graining and track quadratic error energy across shells. Microscopic interactions within each shell cancel identically at the energy level, so shell energies evolve only through dissipation and external inter-shell interactions. We formalize this via a \emph{renormalizable shell-dynamics} assumption, under which cumulative microscopic effects reduce to a controlled net flux across shell boundaries. Assuming an effective power-law spectral transport in a relevant resolution range, the shell dynamics admits a self-similar solution with a moving resolution frontier and explicit scaling exponents. This framework explains neural scaling laws and double descent, and unifies lazy (NTK-like) training and feature learning as two limits of the same spectral-shell dynamics.

Yizhou Zhang, Lun Du

Large-scale neural models are increasingly trained with data pruning, synthetic data generation, cross-model distillation, reinforcement learning from human feedback (RLHF), and difficulty-based sampling. While several of these data-centric strategies reliably improve training efficiency and downstream performance, others fail to provide meaningful gains -- most notably self-generated synthetic data, which often increases dataset volume without enhancing model capability. We formalize data curation as reweighting the sampling distribution and map its effect onto the eigenstructure of the data-induced operator. Our first main result shows that \textbf{static pruning induces a bounded operator and therefore cannot change the spectral tail exponent}; it provides at most finite-region improvements and cannot alter asymptotic neural scaling. Our second result analyzes \textbf{time-dependent data curation}, showing that an ideal oracle capable of tracking spectral residuals and continuously re-normalizing the tail can provably accelerate learning -- although practical systems can only approximate this behavior.

Yizhou Zhang, Yi-An Ma, Eric Mazumdar

We consider the problem of learning to exploit learning algorithms through repeated interactions in games. Specifically, we focus on the case of repeated two player, finite-action games, in which an optimizer aims to steer a no-regret learner to a Stackelberg equilibrium without knowledge of its payoffs. We first show that this is impossible if the optimizer only knows that the learner is using an algorithm from the general class of no-regret algorithms. This suggests that the optimizer requires more information about the learner's objectives or algorithm to successfully exploit them. Building on this intuition, we reduce the problem for the optimizer to that of recovering the learner's payoff structure. We demonstrate the effectiveness of this approach if the learner's algorithm is drawn from a smaller class by analyzing two examples: one where the learner uses an ascent algorithm, and another where the learner uses stochastic mirror ascent with known regularizer and step sizes.

Yizhou Zhang, Yang Sui

This paper explores the intricate behavior of deep neural networks (DNNs) through the lens of neuron activation dynamics. We propose a probabilistic framework that can analyze models' neuron activation patterns as a stochastic process, uncovering theoretical insights into neural scaling laws, such as over-parameterization and the power-law decay of loss with respect to dataset size. By deriving key mathematical relationships, we present that the number of activated neurons increases in the form of $N(1-(\frac{bN}{D+bN})^b)$, and the neuron activation should follows power-law distribution. Based on these two mathematical results, we demonstrate how DNNs maintain generalization capabilities even under over-parameterization, and we elucidate the phase transition phenomenon observed in loss curves as dataset size plotted in log-axis (i.e. the data magnitude increases linearly). Moreover, by combining the above two phenomenons and the power-law distribution of neuron activation, we derived the power-law decay of neural network's loss function as the data size scale increases. Furthermore, our analysis bridges the gap between empirical observations and theoretical underpinnings, offering experimentally testable predictions regarding parameter efficiency and model compressibility. These findings provide a foundation for understanding neural network scaling and present new directions for optimizing DNN performance.

Yizhou Zhang, Karishma Sharma, Yan Liu

Recent years have witnessed an increasing use of coordinated accounts on social media, operated by misinformation campaigns to influence public opinion and manipulate social outcomes. Consequently, there is an urgent need to develop an effective methodology for coordinated group detection to combat the misinformation on social media. However, existing works suffer from various drawbacks, such as, either limited performance due to extreme reliance on predefined signatures of coordination, or instead an inability to address the natural sparsity of account activities on social media with useful prior domain knowledge. Therefore, in this paper, we propose a coordination detection framework incorporating neural temporal point process with prior knowledge such as temporal logic or pre-defined filtering functions. Specifically, when modeling the observed data from social media with neural temporal point process, we jointly learn a Gibbs-like distribution of group assignment based on how consistent an assignment is to (1) the account embedding space and (2) the prior knowledge. To address the challenge that the distribution is hard to be efficiently computed and sampled from, we design a theoretically guaranteed variational inference approach to learn a mean-field approximation for it. Experimental results on a real-world dataset show the effectiveness of our proposed method compared to the SOTA model in both unsupervised and semi-supervised settings. We further apply our model on a COVID-19 Vaccine Tweets dataset. The detection result suggests the presence of suspicious coordinated efforts on spreading misinformation about COVID-19 vaccines.

Karishma Sharma, Yizhou Zhang, Yan Liu

COVID-19 vaccine hesitancy has increased concerns about vaccine uptake required to overcome the pandemic and protect public health. A critical factor associated with anti-vaccine attitudes is the information shared on social media. In this work, we investigate misinformation communities and narratives that can contribute to COVID-19 vaccine hesitancy. During the pandemic, anti-science and political misinformation/conspiracies have been rampant on social media. Therefore, we investigate misinformation and conspiracy groups and their characteristic behaviours in Twitter data collected on COVID-19 vaccines. We identify if any suspicious coordinated efforts are present in promoting vaccine misinformation, and find two suspicious groups - one promoting a 'Great Reset' conspiracy which suggests that the pandemic is orchestrated by world leaders to take control of the economy, with vaccine related misinformation and strong anti-vaccine and anti-social messages such as no lock-downs; and another promoting the Bioweapon theory. Misinformation promoted is largely from the anti-vaccine and far-right communities in the 3-core of the retweet graph, with its tweets proportion of conspiracy and questionable sources to reliable sources being much higher. In comparison with the mainstream and health news, the right-leaning community is more influenced by the anti-vaccine and far-right communities, which is also reflected in the disparate vaccination rates in left and right U.S. states. The misinformation communities are also more vocal, either in vaccine or other discussions, relative to remaining communities, besides other behavioral differences.

Yizhou Zhang, Guojie Song, Lun Du et al.

Recent works reveal that network embedding techniques enable many machine learning models to handle diverse downstream tasks on graph structured data. However, as previous methods usually focus on learning embeddings for a single network, they can not learn representations transferable on multiple networks. Hence, it is important to design a network embedding algorithm that supports downstream model transferring on different networks, known as domain adaptation. In this paper, we propose a novel Domain Adaptive Network Embedding framework, which applies graph convolutional network to learn transferable embeddings. In DANE, nodes from multiple networks are encoded to vectors via a shared set of learnable parameters so that the vectors share an aligned embedding space. The distribution of embeddings on different networks are further aligned by adversarial learning regularization. In addition, DANE's advantage in learning transferable network embedding can be guaranteed theoretically. Extensive experiments reflect that the proposed framework outperforms other state-of-the-art network embedding baselines in cross-network domain adaptation tasks.