arXiv subjects

Zi-Hao Zhou

Publications and source records attributed to Zi-Hao Zhou.

10 recordsLinked to original sources

arXiv2026

Outcome-Guided On-Policy Self-Distillation

On-policy self-distillation (OPSD) provides denser token-level supervision and better computational efficiency than Reinforcement Learning with Verifiable Rewards (RLVR). However, this denser supervision may introduce substantial noise and training instability. Existing improvements often rely on high-variance per-token statistics and introduce extra hyperparameters and trade-offs. Based on the advantage formulation in RLVR, we analyze the OPSD objective from the same perspective, incorporating outcome correctness signals. We find that vanilla OPSD imposes insufficient penalties and excessive rewards on incorrect trajectories because it applies a fixed divergence objective regardless of outcome correctness. Furthermore, the reliability of teacher supervision is associated with both trajectory outcome and the cumulative average teacher entropy along the rollout. Based on these observations, we propose Outcome-Guided On-Policy Self-Distillation (OG-OPSD), which dynamically adapts both the divergence objective and distillation position according to binary outcome rewards and the cumulative average teacher entropy. Extensive experiments show that OG-OPSD consistently improves the performance of vanilla OPSD and multiple strong baselines in mathematical reasoning, multimodal reasoning, and out-of-distribution tasks across Qwen3 models at 1.7B, 4B, and 8B scales, as well as Qwen3-VL-2B.

cs.LG↗

arXiv2026

Learning Process Rewards via Reasoning State Propagation

Process reward models (PRMs) have demonstrated notable effectiveness in test-time scaling and reinforcement learning by providing fine-grained signals for evaluating intermediate reasoning states, but their training relies heavily on costly process annotations. A natural way to alleviate this dependence is to complement limited process supervision with scalable outcome supervision. However, existing PRMs often model reasoning prefixes independently, providing no explicit mechanism for effectively using final outcome to guide the learning of intermediate reasoning states. We introduce Reasoning State Propagation (RSP), which represents each reasoning prefix with a binary validity state and models transitions between successive states across the reasoning trajectory. Specifically, RSP predicts a break probability that a valid state becomes invalid and a repair probability that an invalid state returns to valid. By propagating these transitions, RSP connects intermediate states to the final state, allowing process annotations to supervise intermediate states while outcome labels supervise the final state and can provide learning signals to preceding steps. Across reasoning search, response selection, and reinforcement learning, RSP consistently outperforms representative PRM baselines, with average improvements over Qwen2.5-Math-PRM of 5.6% in beam search and 2.1% in reinforcement learning.

cs.AI↗

arXiv2026

RL Forgets! Towards Continual Policy Optimization

Continual post-training is becoming a central paradigm for adapting vision-language models to evolving tasks. Recent studies report that reinforcement learning is less prone to forgetting than supervised fine-tuning, motivating the view that RL is inherently resistant to forgetting. However, this view remains insufficiently validated, as existing evidence is largely drawn from outdated or homogeneous benchmarks. We revisit this assumption by introducing MRCL, a Multimodal Reasoning Continual Learning benchmark built from recent and diverse multimodal reasoning tasks. Experiments on MRCL show that standard reinforcement learning still suffers from catastrophic forgetting during continual post-training. We trace the failure to an objective mismatch. The KL regularization used in common policy optimization methods is evaluated on current-task data, whereas forgetting is caused by behavioral drift on prior-task distributions. To address this problem, we propose Continual Policy Optimization (CPO), a replay-free method grounded in a prior-task behavioral KL objective. CPO derives a local Fisher surrogate from the historical KL objective and uses parameter movement as a gradient-free proxy for Fisher sensitivity, enabling sparse regularization with negligible additional overhead. Experiments on three model scales and comparisons with multiple RL baselines show that CPO consistently reduces forgetting while maintaining effective adaptation and preserving broader pretrained capabilities. The implementation code is available at https://github.com/MaolinLuo/CPO.

cs.LG↗

arXiv2026

Spectral Imbalance Causes Forgetting in Low-Rank Continual Adaptation

Parameter-efficient continual learning aims to adapt pre-trained models to sequential tasks without forgetting previously acquired knowledge. Most existing approaches treat continual learning as avoiding interference with past updates, rather than considering what properties make the current task-specific update naturally preserve previously acquired knowledge. From a knowledge-decomposition perspective, we observe that low-rank adaptations exhibit highly imbalanced singular value spectra: a few dominant components absorb most of the adaptation energy, thereby (i) more likely to disrupt previously acquired knowledge and (ii) making the update more vulnerable to interference from subsequent tasks. To enable explicit balance among components, we decouple the magnitude of the task update from its directional structure and formulate it as a constrained optimization problem on a restricted Stiefel manifold. We address this problem using a projected first-order method compatible with standard deep-learning optimizers used in vision-language models. Our method mitigates both backward and forward forgetting, consistently outperforming continual learning baselines. The implementation code is available at https://github.com/haodotgu/EBLoRA.

cs.LG↗

arXiv2026

KeepLoRA++: Continual Learning with Layer-Scaled Residual Gradient Adaptation

Continual learning for pre-trained vision-language models requires balancing three competing objectives: retaining pre-trained knowledge, preserving knowledge from a sequence of learned tasks, and maintaining the plasticity to acquire new knowledge. This paper presents KeepLoRA++, balancing these objectives through a unified dual-dimensional knowledge retention mechanism. We analyze knowledge distribution of Transformer architecture from both inter-layer and intra-layer perspectives. The inter-layer perspective examines how retention is distributed across layers, while the intra-layer perspective focuses on the parameter space within each layer. Our analysis reveals a structural property: general transferable knowledge is mainly encoded in the shallow layers and the principal subspace of the parameters, while task-specific adaptations are localized in the deep layers and the residual subspace. Motivated by this insight, KeepLoRA++ introduces a layer-scaled residual gradient adaptation method. New tasks are learned by restricting LoRA parameter updates to the residual subspace, combined with a shallow-to-deep layer scaling, to prevent interference with previously acquired capabilities. Specifically, the gradient of a new task is projected onto a subspace orthogonal to both the principal subspace of the pre-trained model and the dominant directions of previous task features, while simultaneously assigning smaller update magnitudes to shallow layers and larger ones to deeper layers. Our theoretical analysis and empirical evaluations confirm that KeepLoRA++ successfully balances these three competing objectives, consistently outperforming representative baselines across image classification, visual question answering, and video understanding tasks.

cs.CV↗

arXiv2026

DC-Merge: Improving Model Merging with Directional Consistency

Model merging aims to integrate multiple task-adapted models into a unified model that preserves the knowledge of each task. In this paper, we identify that the key to this knowledge retention lies in maintaining the directional consistency of singular spaces between merged multi-task vector and individual task vectors. However, this consistency is frequently compromised by two issues: i) an imbalanced energy distribution within task vectors, where a small fraction of singular values dominate the total energy, leading to the neglect of semantically important but weaker components upon merging, and ii) the geometric inconsistency of task vectors in parameter space, which causes direct merging to distort their underlying directional geometry. To address these challenges, we propose DC-Merge, a method for directional-consistent model merging. It first balances the energy distribution of each task vector by smoothing its singular values, ensuring all knowledge components are adequately represented. These energy-balanced vectors are then projected onto a shared orthogonal subspace to align their directional geometries with minimal reconstruction error. Finally, the aligned vectors are aggregated in the shared orthogonal subspace and projected back to the original parameter space. Extensive experiments on vision and vision-language benchmarks show that DC-Merge consistently achieves state-of-the-art performance in both full fine-tuning and LoRA settings. The implementation code is available at https://github.com/Tobeginwith/DC-Merge.

cs.LG↗

arXiv2026

KeepLoRA: Continual Learning with Residual Gradient Adaptation

cs.CV↗

arXiv2025

LADA: Scalable Label-Specific CLIP Adapter for Continual Learning

Continual learning with vision-language models like CLIP offers a pathway toward scalable machine learning systems by leveraging its transferable representations. Existing CLIP-based methods adapt the pre-trained image encoder by adding multiple sets of learnable parameters, with each task using a partial set of parameters. This requires selecting the expected parameters for input images during inference, which is prone to error that degrades performance. To address this problem, we introduce LADA (Label-specific ADApter). Instead of partitioning parameters across tasks, LADA appends lightweight, label-specific memory units to the frozen CLIP image encoder, enabling discriminative feature generation by aggregating task-agnostic knowledge. To prevent catastrophic forgetting, LADA employs feature distillation for seen classes, preventing their features from being interfered with by new classes. Positioned after the image encoder, LADA prevents gradient flow to the frozen CLIP parameters, ensuring efficient training. Extensive results show that LADA achieves state-of-the-art performance in continual learning settings. The implementation code is available at https://github.com/MaolinLuo/LADA.

cs.CV↗

arXiv2025

Weakly-Supervised Contrastive Learning for Imprecise Class Labels

Contrastive learning has achieved remarkable success in learning effective representations, with supervised contrastive learning often outperforming self-supervised approaches. However, in real-world scenarios, data annotations are often ambiguous or inaccurate, meaning that class labels may not reliably indicate whether two examples belong to the same class. This limitation restricts the applicability of supervised contrastive learning. To address this challenge, we introduce the concept of ``continuous semantic similarity'' to define positive and negative pairs. Instead of directly relying on imprecise class labels, we measure the semantic similarity between example pairs, which quantifies how closely they belong to the same category by iteratively refining weak supervisory signals. Based on this concept, we propose a graph-theoretic framework for weakly-supervised contrastive learning, where semantic similarity serves as the graph weights. Our framework is highly versatile and can be applied to many weakly-supervised learning scenarios. We demonstrate its effectiveness through experiments in two common settings, i.e., noisy label and partial label learning, where existing methods can be easily integrated to significantly improve performance. Theoretically, we establish an error bound for our approach, showing that it can approximate supervised contrastive learning under mild conditions. The implementation code is available at https://github.com/Speechless-10308/WSC.

cs.LG↗

arXiv2024

Continuous Contrastive Learning for Long-Tailed Semi-Supervised Recognition

Long-tailed semi-supervised learning poses a significant challenge in training models with limited labeled data exhibiting a long-tailed label distribution. Current state-of-the-art LTSSL approaches heavily rely on high-quality pseudo-labels for large-scale unlabeled data. However, these methods often neglect the impact of representations learned by the neural network and struggle with real-world unlabeled data, which typically follows a different distribution than labeled data. This paper introduces a novel probabilistic framework that unifies various recent proposals in long-tail learning. Our framework derives the class-balanced contrastive loss through Gaussian kernel density estimation. We introduce a continuous contrastive learning method, CCL, extending our framework to unlabeled data using reliable and smoothed pseudo-labels. By progressively estimating the underlying label distribution and optimizing its alignment with model predictions, we tackle the diverse distribution of unlabeled data in real-world scenarios. Extensive experiments across multiple datasets with varying unlabeled data distributions demonstrate that CCL consistently outperforms prior state-of-the-art methods, achieving over 4% improvement on the ImageNet-127 dataset. Our source code is available at https://github.com/zhouzihao11/CCL

cs.LG↗