Zhan Su

AAAI Conference 2026 Conference Paper

ConvMix: A Mixed-Criteria Data Augmentation Framework for Conversational Dense Retrieval

• Fengran Mo
• Jinghan Zhang
• Yuchen Hui
• Jia Ao Sun
• Zhichao Xu
• Zhan Su
• Jian-Yun Nie

Conversational search aims to satisfy users’ complex information needs via multiple-turn interactions. The key challenge lies in revealing real users’ search intent from the context-dependent queries. Previous studies achieve conversational search by fine-tuning a conversational dense retriever with relevance judgments between pairs of context-dependent queries and documents. However, this training paradigm encounters data scarcity issues. To this end, we propose ConvMix, a mixed-criteria framework to augment conversational dense retrieval, which covers more aspects than existing data augmentation frameworks. We design a two-sided relevance judgment augmentation schema in a scalable manner via the aid of large language models. Besides, we integrate the framework with quality control mechanisms to obtain semantically diverse samples and near-distribution supervisions to combine various annotated data. Experimental results on five widely used benchmarks show that the conversational dense retriever trained by our ConvMix framework outperforms previous baseline methods, which demonstrates our superior effectiveness.

AAAI Conference 2026 Conference Paper

COVR: Collaborative Optimization of VLMs and RL Agent for Visual-Based Control

• Canming Xia
• Peixi Peng
• Guang Tan
• Zhan Su
• Haoran Xu
• Zhenxian Liu
• Luntong Li

Visual reinforcement learning (RL) suffers from poor sample efficiency due to high-dimensional observations in complex tasks. While existing works have shown that vision-language models (VLMs) can assist RL, they often focus on knowledge distillation from the VLM to RL, overlooking the potential of RL-generated interaction data to enhance the VLM. To address this, we propose COVR, a collaborative optimization framework that enables the mutual enhancement of the VLM and RL policies. Specifically, COVR fine-tunes the VLM with RL-generated data to enhance the semantic reasoning ability consistent with the target task, and uses the enhanced VLM to further guide policy learning via action priors. To improve fine-tuning efficiency, we introduce two key modules: (1) an Exploration-Driven Dynamic Filter module that preserves valuable exploration samples using adaptive thresholds based on the degree of exploration, and (2) a Return-Aware Adaptive Loss Weight module that improves the stability of training by quantifying the inconsistency of sampling actions via return signals of RL. We further design a progressive fine-tuning strategy to reduce resource consumption. Extensive experiments show that COVR achieves strong performance across various challenging visual control tasks.

TMLR Journal 2024 Journal Article

Mixture of Latent Experts Using Tensor Products

• Zhan Su
• Fengran Mo
• Prayag Tiwari
• Benyou Wang
• Qiuchi Li
• Jian-Yun Nie
• Jakob Grue Simonsen

In multi-task learning, the conventional approach involves training a model on multiple tasks simultaneously. However, the training signals from different tasks can interfere with one another, potentially leading to \textit{negative transfer}. To mitigate this, we propose a novel \textit{latent-expert} approach (\texttt{TensorPoly}), that balances parameter efficiency with nuanced routing methods. For \textit{experts}, we reparameterize Low-Rank Adaptation (\texttt{LoRA}) by employing an entangled tensor through the use of tensor product operations and name the resulting approach \texttt{TLoRA}. For \textit{routing function}, we tailor two innovative routing functions according to the granularity: \texttt{TensorPoly-I} which directs to each rank within the entangled tensor while \texttt{TensorPoly-II} offers a finer-grained routing approach targeting each order of the entangled tensor. The experimental results from the multi-task T0-benchmark demonstrate that: 1) all latent-expert approaches surpass the corresponding dense approaches, highlighting the potential of modular language models to mitigate negative inference in multi-task learning and deliver superior outcomes. 2) \texttt{TensorPoly-I} achieves higher parameter efficiency in adaptation and outperforms other modular LMs, which shows the potential of our approach in multi-task transfer learning \footnote{The code is released: \url{https://github.com/microsoft/mttl}}.

EAAI Journal 2023 Journal Article

Measuring similarity based on user activeness in recommender systems to improve algorithm scalability

• Jun Ai
• Yifang Cai
• Zhan Su
• Dunlu Peng
• Fengyu Zhao

NeurIPS Conference 2023 Conference Paper

Multi-Head Adapter Routing for Cross-Task Generalization

• Lucas Page-Caccia
• Edoardo Maria Ponti
• Zhan Su
• Matheus Pereira
• Nicolas Le Roux
• Alessandro Sordoni

Parameter-efficient fine-tuning (PEFT) for cross-task generalization consists in pre-training adapters on a multi-task training set before few-shot adaptation to test tasks. Polytropon [Ponti et al. , 2023] ($\texttt{Poly}$) jointly learns an inventory of adapters and a *routing* function that selects a (variable-size) subset of adapters for each task during both pre-training and few-shot adaptation. In this paper, we investigate the role that adapter routing plays in its success and design new variants based on our findings. First, we build on the intuition that finer-grained routing provides more expressivity. Hence, we propose $\texttt{MHR}$ (Multi-Head Routing) which combines *subsets* of adapter parameters and outperforms $\texttt{Poly}$ under a comparable parameter budget; by only fine-tuning the routing function and not the adapters ($\texttt{MHR}$-$z$) we achieve competitive performance with extreme parameter efficiency. Second, we find that $\texttt{Poly}$/$\texttt{MHR}$ performance is a result of better multi-task optimization, rather than modular inductive biases that facilitate adapter recombination and local adaptation, as previously hypothesized. In fact, we find that $\texttt{MHR}$ exhibits high gradient alignment between training tasks. We find that routing is most beneficial during multi-task pre-training rather than during few-shot adaptation and propose $\texttt{MHR}$-$\mu$, which discards routing and fine-tunes the average of the pre-trained adapters on each downstream tasks. This establishes $\texttt{MHR}$-$\mu$ as an effective method for single-adapter fine-tuning. We also show that $\texttt{MHR}$-$\mu$ can be used as an effective zero-shot transfer method by training the average of the pre-trained adapters for a few additional steps on the multi-task training set: this yields gains up to 3\% on absolute accuracy w. r. t. the baselines. Code is available at.

AAAI Conference 2019 Conference Paper

A Generalized Language Model in Tensor Space

• Lipeng Zhang
• Peng Zhang
• Xindian Ma
• Shuqin Gu
• Zhan Su
• Dawei Song

Sentence semantic matching requires an agent to determine the semantic relation between two sentences, which is widely used in various natural language tasks such as Natural Language Inference (NLI) and Paraphrase Identification (PI). Among all matching methods, attention mechanism plays an important role in capturing the semantic relations and properly aligning the elements of two sentences. Previous methods utilized attention mechanism to select important parts of sentences at one time. However, the important parts of the sentence during semantic matching are dynamically changing with the degree of sentence understanding. Selecting the important parts at one time may be insufficient for semantic understanding. To this end, we propose a Dynamic Re-read Network (DRr-Net) approach for sentence semantic matching, which is able to pay close attention to a small region of sentences at each step and re-read the important words for better sentence semantic understanding. To be specific, we first employ Attention Stack-GRU (ASG) unit to model the original sentence repeatedly and preserve all the information from bottom-most word embedding input to up-most recurrent output. Second, we utilize Dynamic Re-read (DRr) unit to pay close attention to one important word at one time with the consideration of learned information and re-read the important words for better sentence semantic understanding. Extensive experiments on three sentence matching benchmark datasets demonstrate that DRr-Net has the ability to model sentence semantic more precisely and significantly improve the performance of sentence semantic matching. In addition, it is very interesting that some of finding in our experiments are consistent with the findings of psychological research.

AAAI Conference 2018 Conference Paper

End-to-End Quantum-like Language Models with Application to Question Answering

• Peng Zhang
• Jiabin Niu
• Zhan Su
• Benyou Wang
• Liqun Ma
• Dawei Song

Language Modeling (LM) is a fundamental research topic in a range of areas. Recently, inspired by quantum theory, a novel Quantum Language Model (QLM) has been proposed for Information Retrieval (IR). In this paper, we aim to broaden the theoretical and practical basis of QLM. We develop a Neural Network based Quantum-like Language Model (NNQLM) and apply it to Question Answering. Specifically, based on word embeddings, we design a new density matrix, which represents a sentence (e. g. , a question or an answer) and encodes a mixture of semantic subspaces. Such a density matrix, together with a joint representation of the question and the answer, can be integrated into neural network architectures (e. g. , 2-dimensional convolutional neural networks). Experiments on the TREC-QA and WIKIQA datasets have verified the effectiveness of our proposed models.