Tianren Zhang

Tianren Zhang, Xiangxin Li, Minghao Xiao et al.

Deep networks often exhibit a preference for "simple" solutions, and such a simplicity bias is widely believed to play a key role in generalization. Yet a broadly applicable, quantitative measure of simplicity remains elusive. We introduce polynomial representations as a distribution-aware, low-dimensional surrogate for neural functions: we approximate a network's predictive behavior along data-dependent interpolation paths using orthogonal polynomial bases, yielding a compact functional representation. We show that the effective degree of this representation serves as a practical simplicity metric that is predictive of generalization across tasks and architectures, and consistently outperforms existing generalization proxies such as sharpness. Finally, polynomial representations naturally yield a differentiable simplicity regularizer, which consistently improves generalization in image and text classification, fine-tuning contrastive vision-language models, and reinforcement learning.

Guanyu Chen, Ruichen Wang, Tianren Zhang et al.

In-context learning (ICL) is a valuable capability exhibited by Transformers pretrained on diverse sequence tasks. However, previous studies have observed that ICL often conflicts with the model's inherent in-weight learning (IWL) ability. By examining the representation space learned by a toy model in synthetic experiments, we identify the shared encoding space for context and samples in Transformers as a potential source of this conflict. To address this, we modify the model architecture to separately encode the context and samples into two distinct spaces: a task representation space and a sample representation space. We model these two spaces under a simple yet principled framework, assuming a linear representational structure and treating them as a pair of dual spaces. Both theoretical analysis and empirical results demonstrate the effectiveness of our proposed architecture, CoQE, in the single-value answer setting. It not only enhances ICL performance through improved representation learning, but also successfully reconciles ICL and IWL capabilities across synthetic few-shot classification and a newly designed pseudo-arithmetic task. Code: https://github.com/McGuinnessChen/dual-representation-space-encoding

Tianrun Xu, Yue Sun, Qixun Wang et al.

Multimodal latent-space reasoning aims to replace explicit thinking with images by performing visual reasoning directly in a compact latent space. However, existing approaches largely rely on visual supervision and produce latent representations that lack sufficient semantic richness, limiting their ability to support diverse region-level reasoning tasks. In this work, we introduce Semantic-Enriched Latent Visual Reasoning (SLVR), a two-stage learning framework that enriches latent representations with attribute-level visual semantics and aligns them with diverse reasoning objectives. In the first stage, SLVR learns semantically enriched region-centric latents under fine-grained attribute supervision. In the second stage, we design Multi-query Group Relative Policy Optimization (M-GRPO) to align latent representations across multiple queries grounded in the same region. To support this framework, we construct SLV-Set, comprising approximately 400K region-level attribute annotations and 800K multi-query question answering samples, and introduce SV-QA, a benchmark that evaluates latent reasoning under semantic variation. Experiments demonstrate that SLVR improves the robustness and semantic consistency of latent visual reasoning compared to existing baselines.

Tianrun Xu, Haoda Jing, Ye Li et al.

Recent advances in multimodal language models (MLLMs) have achieved remarkable progress in vision-language reasoning, especially with the emergence of "thinking with images," which integrates explicit visual steps into the reasoning process. While this paradigm strengthens image-based reasoning, a significant challenge remains: models may arrive at correct answers by relying on irrelevant or spurious regions, driven by prior knowledge or dataset biases. Even when the answer is correct, flawed reasoning indicates that the model has not truly understood the image, highlighting the critical importance of reasoning fidelity in multimodal tasks. To address this issue, we propose DeFacto, a counterfactual reasoning framework that jointly enforces accurate answering and faithful reasoning. A key component of our approach is the design of three complementary training paradigms: (i) positive, (ii) counterfactual, and (iii) random-masking. To enable these paradigms, we develop a pipeline that automatically localizes question-relevant evidence and constructs positive, counterfactual, and random variants, resulting in a dataset of about 100k images. Building on this framework, we train multimodal language models with GRPO-based reinforcement learning, where we design three complementary rewards to guide the model toward accurate answering and evidence-grounded reasoning. Experiments on diverse benchmarks demonstrate that DeFacto substantially improves both answer accuracy and reasoning faithfulness, establishing a stronger foundation for interpretable multimodal reasoning. The code is available on GitHub and the dataset is released on HuggingFace.

Tianren Zhang, Dai-Bei Yang

Contemporary large language models (LLMs), such as GPT-4 and Llama, have harnessed extensive computational power and diverse text corpora to achieve remarkable proficiency in interpreting and generating domain-specific content, including materials science. To leverage the domain knowledge embedded within these models, we propose a simple yet effective multimodal architecture, PolyLLMem, which integrates text embeddings generated by Llama 3 with molecular structure embeddings derived from Uni-Mol, for polymer properties prediction tasks. In our model, Low-rank adaptation (LoRA) layers were also incorporated during the property prediction tasks to refine the embeddings based on our limited polymer dataset, thereby enhancing their chemical relevance for polymer SMILES representation. This balanced fusion of fine-tuned textual and structural information enables PolyLLMem to accurately predict a variety of polymer properties despite the scarcity of training data. Its performance is comparable to, and in some cases exceeds, that of graph-based models, as well as transformer-based models that typically require pretraining on millions of polymer samples. These findings demonstrate that LLM, such as Llama, can effectively capture chemical information encoded in polymer PSMILES, and underscore the efficacy of multimodal fusion of LLM embeddings and molecular structure embeddings in overcoming data scarcity and accelerating the discovery of advanced polymeric materials.

Guanyu Chen, Peiyang Wang, Yizhou Jiang et al.

Large language models (LLMs) have been able to perform various forms of reasoning tasks in a wide range of scenarios, but are they truly engaging in task abstraction and rule-based reasoning beyond mere memorization? To answer this question, we propose a novel experimental approach, Misleading Fine-Tuning (MisFT), to examine whether LLMs perform abstract reasoning by altering their original understanding of fundamental rules. In particular, by constructing datasets with math expressions or logical formulas that contradict correct principles, we fine-tune the model to learn those contradictory rules and assess its generalization ability on unseen test domains. Through a series of experiments, we find that current LLMs are capable of applying contradictory rules to solve practical math word problems and natural language reasoning tasks, implying the presence of an internal mechanism in LLMs that abstracts before reasoning.

Tianren Zhang, Guanyu Chen, Feng Chen

Humans develop world models that capture the underlying generation process of data. Whether neural networks can learn similar world models remains an open problem. In this work, we present the first theoretical results for this problem, showing that in a multi-task setting, models with a low-degree bias provably recover latent data-generating variables under mild assumptions--even if proxy tasks involve complex, non-linear functions of the latents. However, such recovery is sensitive to model architecture. Our analysis leverages Boolean models of task solutions via the Fourier-Walsh transform and introduces new techniques for analyzing invertible Boolean transforms, which may be of independent interest. We illustrate the algorithmic implications of our results and connect them to related research areas, including self-supervised learning, out-of-distribution generalization, and the linear representation hypothesis in large language models.

Tianren Zhang, Chujie Zhao, Guanyu Chen et al.

Learning representations that generalize under distribution shifts is critical for building robust machine learning models. However, despite significant efforts in recent years, algorithmic advances in this direction have been limited. In this work, we seek to understand the fundamental difficulty of out-of-distribution generalization with deep neural networks. We first empirically show that perhaps surprisingly, even allowing a neural network to explicitly fit the representations obtained from a teacher network that can generalize out-of-distribution is insufficient for the generalization of the student network. Then, by a theoretical study of two-layer ReLU networks optimized by stochastic gradient descent (SGD) under a structured feature model, we identify a fundamental yet unexplored feature learning proclivity of neural networks, feature contamination: neural networks can learn uncorrelated features together with predictive features, resulting in generalization failure under distribution shifts. Notably, this mechanism essentially differs from the prevailing narrative in the literature that attributes the generalization failure to spurious correlations. Overall, our results offer new insights into the non-linear feature learning dynamics of neural networks and highlight the necessity of considering inductive biases in out-of-distribution generalization.

Chujie Zhao, Tianren Zhang, Feng Chen

Domain generalization (DG) aims to improve the generalization ability of the model trained on several known training domains over unseen test domains. Previous work has shown that self-supervised contrastive pre-training improves the robustness of the model on downstream tasks. However, in this paper, we find that self-supervised models do not exhibit better generalization performance than supervised models pre-trained on the same dataset in the DG setting. We argue that this is owing to the fact that the richer intra-class discriminative features extracted by self-supervised contrastive learning, which we term silent features, are suppressed during supervised fine-tuning. These silent features are likely to contain features that are more generalizable on the test domain. In this work, we model and analyze this feature suppression phenomenon and theoretically prove that preserving silent features can achieve lower expected test domain risk under certain conditions. In light of this, we propose a simple yet effective method termed STEP (Silent Feature Preservation) to improve the generalization performance of the self-supervised contrastive learning pre-trained model by alleviating the suppression of silent features during the supervised fine-tuning process. Experimental results show that STEP exhibits state-of-the-art performance on standard DG benchmarks with significant distribution shifts.

Yihan Li, Jinsheng Ren, Tianrun Xu et al.

Recently, incorporating natural language instructions into reinforcement learning (RL) to learn semantically meaningful representations and foster generalization has caught many concerns. However, the semantical information in language instructions is usually entangled with task-specific state information, which hampers the learning of semantically invariant and reusable representations. In this paper, we propose a method to learn such representations called element randomization, which extracts task-relevant but environment-agnostic semantics from instructions using a set of environments with randomized elements, e.g., topological structures or textures, yet the same language instruction. We theoretically prove the feasibility of learning semantically invariant representations through randomization. In practice, we accordingly develop a hierarchy of policies, where a high-level policy is designed to modulate the behavior of a goal-conditioned low-level policy by proposing subgoals as semantically invariant representations. Experiments on challenging long-horizon tasks show that (1) our low-level policy reliably generalizes to tasks against environment changes; (2) our hierarchical policy exhibits extensible generalization in unseen new tasks that can be decomposed into several solvable sub-tasks; and (3) by storing and replaying language trajectories as succinct policy representations, the agent can complete tasks in a one-shot fashion, i.e., once one successful trajectory has been attained.

Tianren Zhang, Shangqi Guo, Tian Tan et al.

Goal-conditioned Hierarchical Reinforcement Learning (HRL) is a promising approach for scaling up reinforcement learning (RL) techniques. However, it often suffers from training inefficiency as the action space of the high-level, i.e., the goal space, is large. Searching in a large goal space poses difficulty for both high-level subgoal generation and low-level policy learning. In this paper, we show that this problem can be effectively alleviated by restricting the high-level action space from the whole goal space to a $k$-step adjacent region of the current state using an adjacency constraint. We theoretically prove that in a deterministic Markov Decision Process (MDP), the proposed adjacency constraint preserves the optimal hierarchical policy, while in a stochastic MDP the adjacency constraint induces a bounded state-value suboptimality determined by the MDP's transition structure. We further show that this constraint can be practically implemented by training an adjacency network that can discriminate between adjacent and non-adjacent subgoals. Experimental results on discrete and continuous control tasks including challenging simulated robot locomotion and manipulation tasks show that incorporating the adjacency constraint significantly boosts the performance of state-of-the-art goal-conditioned HRL approaches.

Tianren Zhang, Yizhou Jiang, Feng Chen

Conventional supervised learning assumes a stable input-output relationship. However, this assumption fails in open-ended training settings where the input-output relationship depends on hidden contexts. In this work, we formulate a more general supervised learning problem in which training data is drawn from multiple unobservable domains, each potentially exhibiting distinct input-output maps. This inherent conflict in data renders standard empirical risk minimization training ineffective. To address this challenge, we propose a method LEAF that introduces an allocation function, which learns to assign conflicting data to different predictive models. We establish a connection between LEAF and a variant of the Expectation-Maximization algorithm, allowing us to derive an analytical expression for the allocation function. Finally, we provide a theoretical analysis of LEAF and empirically validate its effectiveness on both synthetic and real-world tasks involving conflicting data.

Chongkai Gao, Haichuan Gao, Shangqi Guo et al.

Imitation learning (IL) algorithms have shown promising results for robots to learn skills from expert demonstrations. However, they need multi-task demonstrations to be provided at once for acquiring diverse skills, which is difficult in real world. In this work we study how to realize continual imitation learning ability that empowers robots to continually learn new tasks one by one, thus reducing the burden of multi-task IL and accelerating the process of new task learning at the same time. We propose a novel trajectory generation model that employs both a generative adversarial network and a dynamics-aware prediction model to generate pseudo trajectories from all learned tasks in the new task learning process. Our experiments on both simulation and real-world manipulation tasks demonstrate the effectiveness of our method.

Tianren Zhang, Shangqi Guo, Tian Tan et al.

Goal-conditioned hierarchical reinforcement learning (HRL) is a promising approach for scaling up reinforcement learning (RL) techniques. However, it often suffers from training inefficiency as the action space of the high-level, i.e., the goal space, is often large. Searching in a large goal space poses difficulties for both high-level subgoal generation and low-level policy learning. In this paper, we show that this problem can be effectively alleviated by restricting the high-level action space from the whole goal space to a $k$-step adjacent region of the current state using an adjacency constraint. We theoretically prove that the proposed adjacency constraint preserves the optimal hierarchical policy in deterministic MDPs, and show that this constraint can be practically implemented by training an adjacency network that can discriminate between adjacent and non-adjacent subgoals. Experimental results on discrete and continuous control tasks show that incorporating the adjacency constraint improves the performance of state-of-the-art HRL approaches in both deterministic and stochastic environments.