Sequential Recommender Systems (SRSs) predict items of interest for users based on their historical interactions. Two key popular paradigms for SRs are unidirectional Next Item Prediction (NIP) and bidirectional Masked Language Modeling (MLM). NIP performs well in recommendation tasks but is constrained by its reliance on prior information and a rigid temporal order assumption, limiting its ability to capture dynamic personalized preferences. MLM, on the other hand, enhances sequence representation and personalization by leveraging bidirectional context, but its objective is misaligned with recommendation tasks. To this end, we develop MetaSR, a meta-learning-based SR approach that ingeniously combines both paradigms into a unified framework to enhance personalized sequential recommendation. Specifically, we treat the mask prediction task in MLM as the inner-loop task (i.e., the support set), adjusting mask items dynamically to generate sequences that match individual behavioral patterns to capture personalized user preferences. In the outer loop (i.e., the query set), we adopt the NIP paradigm to directly train MetaSR toward the SR target. The joint training bridges the gap between MLM and SRs tasks, also overcoming NIP’s limitations in capturing dynamic personalized preferences. Furthermore, to improve the overall efficiency, we develop a reinforcement learning-based Adaptive Masked Sequence Selection (AMSS) mechanism, automatically selecting the optimal masking prediction tasks within the meta-learning support set to accelerate the meta-learning and achieve better personalization. Experiments on multiple public benchmark datasets show that MetaSR significantly outperforms existing SRs models. The dataset and codes are available at MetaSR .

Automated identification of bacterial flagellar motors in cryo-electron tomography (cryo-ET) remains a significant challenge due to low signal-to-noise ratios, imaging artifacts, and the complexity of capturing both local and global structural contexts. In this work, we propose MSVT+AUM, a hybrid framework that combines a Multi-Scale Volumetric Transformer (MSVT) with Adaptive Uncertainty Modeling (AUM) to address these limitations. MSVT employs a 3D ResNet-50 backbone to extract hierarchical volumetric features, which are integrated through a volume-level self-attention mechanism and refined via 3D Vision Transformer blocks with positional encoding. To enhance discriminative capability in densely packed molecular environments, we introduce a Contextual Spatial Attention module that adaptively reweights spatial and channel-wise representations. To reduce false positives and provide calibrated confidence estimates, AUM jointly models aleatoric and epistemic uncertainties using heteroscedastic regression and Monte Carlo dropout, with detection thresholds dynamically adjusted based on uncertainty estimates. Evaluation on the BYU flagellar motor dataset and an extended subset of the CryoET Data Portal shows that our method achieves an F 2 score of 0.891, representing a 1.5 % improvement over existing approaches, while reducing in-ference time by 23%. MSVT+AUM exhibits robust generalization across diverse bacterial species and noise conditions, offering a scalable and reliable solution for high-throughput structural analysis in cryo-ET.

Knowledge Tracing (KT) is a fundamental task in personalized education, aiming to predict student performance by modeling their evolving concept mastery. Recent state-of-the-art approaches adopt multi-graph architectures to capture diverse concept and behavior relations. However, such models often suffer from graph imbalance, where one graph branch dominates training, undermining the benefits of structural integration. To address this, we propose AMBER (Adaptive Meta-Balanced Ensemble Representation learning), a KT framework designed to promote balanced learning across heterogeneous graph structures. AMBER introduces an external dual-graph teacher to guide the learning of ensemble representations. As the teacher itself may encode graph imbalance bias, we further incorporate a meta-distillation strategy that adaptively adjusts the teacher using student feedback, amplifying signals beneficial to underperforming branches. In addition, an adaptive graph rebalancing strategy is introduced to balance the optimization of different graph branches in real time, preventing dominance by any single structure. Experiments on three real-world datasets show that AMBER consistently outperforms competitive baselines. By promoting balanced optimization across graphs, AMBER enables more effective integration of heterogeneous learning signals in KT, providing a robust and scalable solution for personalized education. Code is available at https://github.com/AMBER2025KT/AMBER2025CIKM.

Creating 3D content from single-view images is a challenging problem that has attracted considerable attention in recent years. Current approaches typically utilize score distillation sampling (SDS) from pre-trained 2D diffusion models to generate multi-view 3D representations. Although some methods have made notable progress by balancing generation speed and model quality, their performance is often limited by the visual inconsistencies of the diffusion model outputs. In this work, we propose ContrastiveGaussian, which integrates contrastive learning into the generative process. By using a perceptual loss, we effectively differentiate between positive and negative samples, leveraging the visual inconsistencies to improve 3D generation quality. To further enhance sample differentiation and improve contrastive learning, we incorporate a super-resolution model and introduce another Quantity-Aware Triplet Loss to address varying sample distributions during training. Our experiments demonstrate that our approach achieves superior texture fidelity and improved geometric consistency. Code will be available at https://github.com/YaNLlan-Ljb/ContrastiveGaussian.

Modern text-to-image generation systems have enabled the creation of remarkably realistic and high-quality visuals, yet they often falter when handling the inherent ambiguities in user prompts. In this work, we present Twin-Co, a framework that leverages synchronized, co-adaptive dialogue to progressively refine image generation. Instead of a static generation process, Twin-Co employs a dynamic, iterative workflow where an intelligent dialogue agent continuously interacts with the user. Initially, a base image is generated from the user's prompt. Then, through a series of synchronized dialogue exchanges, the system adapts and optimizes the image according to evolving user feedback. The co-adaptive process allows the system to progressively narrow down ambiguities and better align with user intent. Experiments demonstrate that Twin-Co not only enhances user experience by reducing trial-and-error iterations but also improves the quality of the generated images, streamlining creative process across various applications.

Due to the high cost and small scale of Image Quality Assessment (IQA) datasets, achieving robust generalization remains challenging for prevalent Blind IQA (BIQA) methods. Traditional deep learning-based methods emphasize visual information to capture quality features, while recent developments in Vision-Language Models (VLMs) demonstrate strong potential in learning generalizable representations through textual information. However, applying VLMs to BIQA poses three major Challenges: (1) How to make full use of the multi-modal information. (2) The prompt engineering for appropriate quality description is extremely time-consuming. (3) How to use mixed data for joint training to enhance the generalization of VLM-based BIQA model. To this end, we propose a Multi-modal BIQA method with prompt learning, named MMP-IQA. For (1), we propose a conditional fusion module to better utilize the cross-modality information. By jointly adjusting visual and textual features, our model can capture quality information with a stronger representation ability. For (2), we model the quality prompt's context words with learnable vectors during the training process, which can be adaptively updated for superior performances. For (3), we jointly train a linearity-induced quality evaluator, a relative quality evaluator, and a dataset-specific absolute quality evaluator. In addition, we propose a dual automatic weight adjustment strategy to adaptively balance the loss weights between different datasets and among various losses within the same dataset. Extensive experiments illustrate the superior effectiveness of MMP-IQA.

Modern Large Language Models, such as the LLaMA, Qwen and DeepSeek series, predominantly adopt the Pre-LayerNorm (Pre-LN) Transformer architecture. While being stable during pretraining and scalable to large model sizes, Pre-LN suffers from an exponential growth in activation variance across layers, causing the shortcut to dominate over sub-layer outputs in the residual connection and limiting the learning capacity of deeper layers. To mitigate this issue, we propose Gradient-Preserving Activation Scaling (GPAS), a simple technique that can be used in combination with existing approaches. GPAS works by scaling down the intermediate activations while keeping their gradients unchanged. This leaves information in the activations intact, and avoids the gradient vanishing problem associated with gradient downscaling. Extensive experiments across various model sizes from 71M to 1B show that GPAS achieves consistent performance gains. Beyond enhancing Pre-LN Transformers, GPAS also shows promise in improving alternative architectures such as Sandwich-LN and DeepNorm, demonstrating its versatility and potential for improving training dynamics in a wide range of settings. Our code is available at https://github.com/dandingsky/GPAS.

We introduce TRiCo, a novel triadic game-theoretic co-training framework that rethinks the structure of semi-supervised learning by incorporating a teacher, two students, and an adversarial generator into a unified training paradigm. Unlike existing co-training or teacher-student approaches, TRiCo formulates SSL as a structured interaction among three roles: (i) two student classifiers trained on frozen, complementary representations, (ii) a meta-learned teacher that adaptively regulates pseudo-label selection and loss balancing via validation-based feedback, and (iii) a non-parametric generator that perturbs embeddings to uncover decision boundary weaknesses. Pseudo-labels are selected based on mutual information rather than confidence, providing a more robust measure of epistemic uncertainty. This triadic interaction is formalized as a Stackelberg game, where the teacher leads strategy optimization and students follow under adversarial perturbations. By addressing key limitations in existing SSL frameworks—such as static view interactions, unreliable pseudo-labels, and lack of hard sample modeling—TRiCo provides a principled and generalizable solution. Extensive experiments on CIFAR-10, SVHN, STL-10, and ImageNet demonstrate that TRiCo consistently achieves state-of-the-art performance in low-label regimes, while remaining architecture-agnostic and compatible with frozen vision backbones.

In this paper, we re-introduce the Curse of Depth, a concept that re-introduces, explains, and addresses the recent observation in modern Large Language Models (LLMs) where deeper layers are much less effective than expected. We first confirm the wide existence of this phenomenon across the most popular families of LLMs, such as Llama, Mistral, DeepSeek, and Qwen. Our analysis, theoretically and empirically, identifies that the underlying reason for the ineffectiveness of deep layers in LLMs is the widespread usage of Pre-Layer Normalization (Pre-LN). While Pre-LN stabilizes the training of Transformer LLMs, its output variance exponentially grows with the model depth, which undesirably causes the derivative of the deep Transformer blocks to be an identity matrix, and therefore barely contributes to the training. To resolve this training pitfall, we propose LayerNorm Scaling, which scales the variance of output of the layer normalization inversely by the square root of its depth. This simple modification mitigates the output variance explosion of deeper Transformer layers, improving their contribution. Our experimental results, spanning model sizes from 130M to 7B, demonstrate that \ours significantly enhances LLM pre-training performance compared to Pre-LN. Moreover, this improvement seamlessly carries over to supervised fine-tuning. All these gains can be attributed to the fact that LayerNorm Scaling enables deeper layers to contribute more effectively during training.