Specializing Large Language Models for educational domains is a key frontier in creating personalized learning tools. The central challenge is not data scarcity but its abundance: efficiently selecting a curated data subset from vast corpora to enhance specialized skills and foster generalization, without degrading existing abilities. Existing data selection paradigms, relying on superficial semantic similarity or model training dynamics, often lack a principled framework to identify data that promotes true cognitive growth. Our work proposes a paradigm shift from leveraging indirect proxies of learning value, such as semantic similarity and training dynamics, towards a framework that performs a direct, cognitive-level modeling of the learner's state. We introduce CASS, a novel framework that implements this cognitive approach through a clear pipeline, moving from an initial Diagnosis to the ultimate goal of expanding the model's cognitive frontier. First, CASS diagnoses the LLM's cognitive frontier using Multidimensional Item Response Theory. Leveraging this diagnosis, it then employs Fisher Information to select a data subset situated at LLM's cognitive frontier that offers maximum informational gain. Finally, the model is fine-tuned on this curated data using a structured, easy-to-hard curriculum to ensure effective learning. Experiments on our new multi-subject dataset show that models trained with CASS not only achieve superior accuracy in the target domain but also exhibit enhanced generalization. CASS provides a more efficient, effective, and theoretically-grounded paradigm for building expert educational LLMs.

Recent studies indicate a fundamental incompatibility between ID representations and language model (LM) representations, as they capture behavioral and semantic spaces respectively.This mismatch leads LM representations to consistently underperform ID representations in recommendation tasks.In this work, we revisit this problem and show, from an information-theoretic perspective, that LLM representations retain all discriminative information in ID representations.Based on this, we introduce a Profile-then-Embedding (PtE) framework for recommendation, consisting of a Profile Stage, in which semantic user and item profiles are generated jointly through LLM-based bidirectional reasoning over useritem interactions, and a Personalized Embedding Stage, which encodes these profiles into task-aligned recommendation embeddings.We demonstrate PtE's effectiveness across three benchmark datasets, including cold-start and long-tail scenarios, achieving substantial gains in both discriminative and generative recommendation models.

Multimodal Document Question Answering (MDQA) systems commonly follow a retrieve-then-answer paradigm; however, the retrieval stage often introduces substantial noise, making an effective reranking component indispensable. Existing reranker training frameworks in MDQA rely predominantly on proxy supervision derived from human annotations or large language model (LLM) outputs, which are frequently noisy and, more critically, misaligned with downstream answer quality. To overcome this limitation, we propose AD-Reranker, a novel framework that shifts reranker training from proxy imitation to answer-driven utility optimization. Specifically, we reformulate the reranker as an environment-grounded agent that interacts with a downstream reader, modeled as a deterministic environment. We further design a composite reward function that integrates answer correctness, thereby explicitly incentivizing ranking strategies aligned with downstream task performance. To optimize the agent, we adopt Group Relative Policy Optimization (GRPO), enabling stable and effective group-wise policy learning. Empirical results demonstrate that AD-Reranker achieves superior reranking quality and an optimal accuracy-efficiency trade-off. When integrated into standard MDQA pipelines, AD-Reranker improves accuracy by 1.9%–5.0% while reducing the reader's context token consumption by 15%–52%, providing strong evidence for the effectiveness of answer-driven reranker training.

Implicit feedback is central to modern recommender systems but is inherently noisy, often impairing model training and degrading user experience. At scale, such noise can mislead learning processes, reducing both recommendation accuracy and platform value. Existing denoising strategies typically overlook the entity-specific nature of noise while introducing high computational costs and complex hyperparameter tuning. To address these challenges, we propose EARD (Entity-Aware Reliability-Driven Denoising), a lightweight framework that shifts the focus from interaction-level signals to entity-level reliability. Motivated by the empirical observation that training loss correlates with noise, EARD quantifies user and item reliability via their average training losses as a proxy for reputation, and integrates these entity-level factors with interaction-level confidence. The framework is model-agnostic, computationally efficient, and requires only two intuitive hyperparameters. Extensive experiments across multiple datasets and backbone models demonstrate that EARD yields substantial improvements over state-of-the-art baselines (e.g., up to 27.01% gain in NDCG@50), while incurring negligible additional computational cost. Comprehensive ablation studies and mechanism analyses further confirm EARD's robustness to hyperparameter choices and its practical scalability. These results highlight the importance of entity-aware reliability modeling for denoising implicit feedback and pave the way for more robust recommendation research.

Large-scale online marketplaces and recommender systems are crucial technological foundations for the development of e-commerce. In industrial recommender systems, features play a vital role as they carry essential information for downstream models. Accurate estimation of feature importance is critical, as it helps identify the most useful feature subsets from thousands of candidates for online services. Such a selection enables optimization of online performance while reducing computational burden. To address the feature selection challenges in deep learning, trainable gate-based and sensitivity-based methods have been proposed and proven effective in the industry. However, by analyzing real-world examples, we identified three bias issues that cause feature importance estimation to rely on partial model layers, samples, or gradients, ultimately leading to inaccurate feature importance estimates. We refer to these biases as layer bias, baseline bias, and approximation bias. To mitigate these biases, we propose FairFS, a fair and accurate feature selection algorithm. On one hand, FairFS directly regularizes feature importance estimation across all non-linear transformational layers to avoid layer bias. On the other hand, it employs a smooth baseline feature close to the classifier's decision boundary and an aggregated approximation method to mitigate bias issues. Extensive experiments demonstrate how FairFS mitigates these three biases and achieves state-of-the-art feature selection results.

Large language models (LLMs) have demonstrated exceptional error detection capabilities and can correct sentences with high fluency in grammatical error correction (GEC) tasks. However, when correcting Chinese academic papers, LLMs face significant challenges of over-correction. To delve deeper into this issue, we explore the underlying reasons. On one hand, each discipline has its unique vocabulary and expressions, and LLMs have insufficient and incomplete understanding of domain-specific sentences. On the other hand, the controllability of generative LLMs in GEC tasks is inherently poor, and the traditional sequence-to-sequence (Seq2Seq) correction structure exacerbates this issue. Considering the two aforementioned factors, we propose a new error correction framework for Chinese academic GEC tasks using LLMs, named ScholarGEC. To improve LLMs’ understanding of domain-specific knowledge, we construct appropriate disciplinary knowledge prefixes for sentences and use this domain-specific knowledge data to fine-tune the LLM. To enhance the controllability of LLMs, we replace the traditional Seq2Seq structure with a Detection-Correction separated structure. We also introduce a special token during the process to improve the model’s error detection stability. Additionally, we incorporate iterative self-reflection to enhance the stability of the generation, in the three parts of LLM generation. Extensive experiments demonstrate the effectiveness and robustness of our framework on a Chinese GEC dataset composed of academic papers, and further analysis reveals the capabilities of our framework in enhancing LLM performance in general GEC tasks.

Although previous studies have applied diffusion models to time series forecasting, these efforts have struggled to preserve the intrinsic temporal correlations within the series, leading to suboptimal predictive outcomes. This failure primarily results from the introduction of independent, identically distributed (i.i.d.) noise. In the forward process, the addition of i.i.d. noise to the time series gradually diminishes these temporal correlations. The reverse process starts with i.i.d. noise and lacks priors related to temporal correlations, which can result in directional biases during sampling. From a frequency-domain perspective, noise disrupts the low-frequency-dominated structure of trend components, making it difficult for the model to learn long-term temporal dependencies. To address these limitations, we introduce a decomposition prediction framework to complement the novel Temporal Correlation-Empowered Diffusion Model. Overall, We decompose the time series into trend and residual components, predict them using a base model and a diffusion model, and then combine the results. Specifically, a frequency-domain MLP model was adopted as the base model due to its not distorting the original sequence, and better the capture of long-range temporal dependencies. The diffusion model incorporates two key modules to capture short- and mid-range temporal correlations: the Maintaining Temporal Correlation Module and the Redesigned Initial Module. Extensive experiments across multiple datasets demonstrate that the proposed method significantly outperforms related strong baselines.

Feature interaction modeling is crucial for deep recommendation models. A common and effective approach is to construct explicit feature combinations to enhance model performance. However, in practice, only a small fraction of these combinations are truly informative. Thus it is essential to select useful feature combinations to reduce noise and manage memory consumption. While feature selection methods have been extensively studied, they are typically limited to selecting individual features. Extending these methods for high-order feature combination selection presents a significant challenge due to the exponential growth in time complexity when evaluating feature combinations one by one. In this paper, we propose TayFCS, a lightweight feature combination selection method that significantly improves model performance. Specifically, we propose the Taylor Expansion Scorer (TayScorer) module for field-wise Taylor expansion on the base model. Instead of evaluating all potential feature combinations' importance by repeatedly running experiments with feature adding and removal, this scorer only needs to approximate them based on their sub-components' gradients. They can be simply computed with one backward pass based on a trained recommendation model. To further reduce information redundancy between feature combinations and their sub-components, we introduce Logistic Regression Elimination (LRE) that estimates the information gain of feature combinations over their sub-components based on the above importance scores. Experimental results on three benchmark datasets validate both the effectiveness and efficiency of our approach. Furthermore, online A/B test results demonstrate its practical applicability and commercial value.

In the realm of recommender systems, accurately predicting Click-Through Rate (CTR) is a critical task that involves learning user-item interaction features. Many researchers propose novel models to mine interaction signals, but they neglect that redundancy itself causes high computational cost and leads to suboptimal performance. Some tried to remove redundancy by dropping useless features, or shrinking the size of embedding table. However, current feature selection methods are vulnerable to training stochasticity and data dynamics, while embedding size assignment techniques neglect the importance relationships between feature fields. The simple combination of the two optimization ways will also yield poor performance due to the inherent gap in their optimization targets. Hence, there is no effective paradigm that can optimize feature fields from the two aspects in a simultaneous and coordinated way. In this paper, we identify the core issue as the lack of a practical score to measure the contribution of feature fields, and propose a distribution-based field optimization framework that adopts importance distribution to provide a comprehensive view for both methods. We innovatively design a learner for each field to acquire the stable and comprehensive importance situation. Then, based on this, we eliminate noise features, and assign adaptive embedding sizes for different feature fields according to the similarity of importance. With this field optimization, our proposed framework has extremely low pre-training overhead, greatly reduces training and inference time, and even achieves more accurate prediction results with fewer feature fields.

With the rapid progress of large language models (LLMs) and diffusion models, there has been growing interest in personalized content generation. However, current conversational systems often present the same recommended content to all users, falling into the dilemma of "one-size-fits-all." To break this limitation and boost user engagement, in this paper, we introduce PCG (**P**ersonalized Visual **C**ontent **G**eneration), a unified framework for personalizing item images within conversational systems. We tackle two key bottlenecks: the depth of personalization and the fidelity of generated images. Specifically, an LLM-powered Inclinations Analyzer is adopted to capture user likes and dislikes from context to construct personalized prompts. Moreover, we design a dual-stage LoRA mechanism—Global LoRA for understanding task-specific visual style, and Local LoRA for capturing preferred visual elements from conversation history. During training, we introduce the visual content condition method to ensure LoRA learns both historical visual context and maintains fidelity to the original item images. Extensive experiments on benchmark conversational datasets—including objective metrics and GPT-based evaluations—demonstrate that our framework outperforms strong baselines, which highlight its potential to redefine personalization in visual content generation for conversational scenarios like e-commerce and real-world recommendation.

Aspect-based sentiment analysis (ABSA) aims to predict the sentiment polarity of a specific aspect within a given sentence.Most existing methods predominantly leverage semantic or syntactic information based on attention scores, which are susceptible to interference caused by irrelevant contexts and often lack sentiment knowledge at a data-specific level.In this paper, we propose a novel Dynamic Multigranularity Attribution Network (DMAN) from the perspective of attribution.Initially, we leverage Integrated Gradients to dynamically extract attribution scores for each token, which contain underlying reasoning knowledge for sentiment analysis.Subsequently, we aggregate attribution representations from multiple semantic granularities in natural language, enhancing a profound understanding of the semantics.Finally, we integrate attribution scores with syntactic information to capture the relationships between aspects and their relevant contexts more accurately during the sentence understanding process.Extensive experiments on five benchmark datasets demonstrate the effectiveness of our proposed method.

The emergence of personalized generation has made it possible to create texts or images that meet the unique needs of users.Recent advances mainly focus on style or scene transfer based on given keywords.However, in e-commerce and recommender systems, it is almost an untouched area to explore user historical interactions, automatically mine user interests with semantic associations, and create item representations that closely align with user individual interests.In this paper, we propose a brand new framework called Interest Augmented Multimodal Generator (I-AM-G).The framework first extracts tags from the multimodal information of items that the user has interacted with, and the most frequently occurred ones are extracted to rewrite the text description of the item.Then, the framework uses a decoupled text-to-text and image-to-image retriever to search for the top-K similar item text and image embeddings from the item pool.Finally, the Attention module for user interests fuses the retrieved information in a cross-modal manner and further guides the personalized generation process collaborating with the rewritten text.We conducted extensive and comprehensive experiments to demonstrate that our framework can effectively generate results aligned with user preferences, which potentially provides a new paradigm of Rewrite and Retrieve for personalized generation.