Tong Xu

Junyi Zhu, Shuochen Liu, Yu Yu et al.

Large language models (LLMs) excel in generating coherent text, but they often struggle with context awareness, leading to inaccuracies in tasks requiring faithful adherence to provided information. We introduce FastMem, a novel method designed to enhance instruction fine-tuned LLMs' context awareness through fast memorization of the prompt. FastMem maximizes the likelihood of the prompt before inference by updating only the last Feed-Forward Network (FFN) module. This targeted approach ensures efficient optimization without overfitting, significantly improving the model's ability to comprehend and accurately follow the context. Our experiments demonstrate substantial gains in reading comprehension, text summarization and adherence to output structures. For instance, FastMem improves the accuracy of Llama 3-8B-Inst on the NQ-SWAP dataset from 59.1% to 71.6%, and reduces the output structure failure rate of Qwen 1.5-4B-Chat from 34.9% to 25.5%. Extensive experimental results highlight FastMem's potential to offer a robust solution to enhance the reliability and accuracy of LLMs in various applications. Our code is available at: https://github.com/IAAR-Shanghai/FastMem

Linlang Jiang, Jingbo Zhou, Tong Xu et al.

Trip recommender system, which targets at recommending a trip consisting of several ordered Points of Interest (POIs), has long been treated as an important application for many location-based services. Currently, most prior arts generate trips following pre-defined objectives based on constraint programming, which may fail to reflect the complex latent patterns hidden in the human mobility data. And most of these methods are usually difficult to respond in real time when the number of POIs is large. To that end, we propose an Adversarial Neural Trip Recommendation (ANT) framework to tackle the above challenges. First of all, we devise a novel attention-based encoder-decoder trip generator that can learn the correlations among POIs and generate well-designed trips under given constraints. Another novelty of ANT relies on an adversarial learning strategy integrating with reinforcement learning to guide the trip generator to produce high-quality trips. For this purpose, we introduce a discriminator, which distinguishes the generated trips from real-life trips taken by users, to provide reward signals to optimize the generator. Moreover, we devise a novel pre-train schema based on learning from demonstration, which speeds up the convergence to achieve a sufficient-and-efficient training process. Extensive experiments on four real-world datasets validate the effectiveness and efficiency of our proposed ANT framework, which demonstrates that ANT could remarkably outperform the state-of-the-art baselines with short response time.

Yanan Wang, Yong Ge, Li Li et al.

Reinforcement learning (RL) has shown great promise in optimizing long-term user interest in recommender systems. However, existing RL-based recommendation methods need a large number of interactions for each user to learn a robust recommendation policy. The challenge becomes more critical when recommending to new users who have a limited number of interactions. To that end, in this paper, we address the cold-start challenge in the RL-based recommender systems by proposing a meta-level model-based reinforcement learning approach for fast user adaptation. In our approach, we learn to infer each user's preference with a user context variable that enables recommendation systems to better adapt to new users with few interactions. To improve adaptation efficiency, we learn to recover the user policy and reward from only a few interactions via an inverse reinforcement learning method to assist a meta-level recommendation agent. Moreover, we model the interaction relationship between the user model and recommendation agent from an information-theoretic perspective. Empirical results show the effectiveness of the proposed method when adapting to new users with only a single interaction sequence. We further provide a theoretical analysis of the recommendation performance bound.

Bin He, Di Zhou, Jinghui Xiao et al.

Complex node interactions are common in knowledge graphs, and these interactions also contain rich knowledge information. However, traditional methods usually treat a triple as a training unit during the knowledge representation learning (KRL) procedure, neglecting contextualized information of the nodes in knowledge graphs (KGs). We generalize the modeling object to a very general form, which theoretically supports any subgraph extracted from the knowledge graph, and these subgraphs are fed into a novel transformer-based model to learn the knowledge embeddings. To broaden usage scenarios of knowledge, pre-trained language models are utilized to build a model that incorporates the learned knowledge representations. Experimental results demonstrate that our model achieves the state-of-the-art performance on several medical NLP tasks, and improvement above TransE indicates that our KRL method captures the graph contextualized information effectively.

Hanqing Tao, Shiwei Tong, Tong Xu et al.

Recent studies have consistently given positive hints that morphology is helpful in enriching word embeddings. In this paper, we argue that Chinese word embeddings can be substantially enriched by the morphological information hidden in characters which is reflected not only in strokes order sequentially, but also in character glyphs spatially. Then, we propose a novel Dual-channel Word Embedding (DWE) model to realize the joint learning of sequential and spatial information of characters. Through the evaluation on both word similarity and word analogy tasks, our model shows its rationality and superiority in modelling the morphology of Chinese.

Zhirui Zhang, Shuangzhi Wu, Shujie Liu et al.

Although Neural Machine Translation (NMT) has achieved remarkable progress in the past several years, most NMT systems still suffer from a fundamental shortcoming as in other sequence generation tasks: errors made early in generation process are fed as inputs to the model and can be quickly amplified, harming subsequent sequence generation. To address this issue, we propose a novel model regularization method for NMT training, which aims to improve the agreement between translations generated by left-to-right (L2R) and right-to-left (R2L) NMT decoders. This goal is achieved by introducing two Kullback-Leibler divergence regularization terms into the NMT training objective to reduce the mismatch between output probabilities of L2R and R2L models. In addition, we also employ a joint training strategy to allow L2R and R2L models to improve each other in an interactive update process. Experimental results show that our proposed method significantly outperforms state-of-the-art baselines on Chinese-English and English-German translation tasks.