Jiayi Zhang

Xiuying Chen, Mingzhe Li, Jiayi Zhang et al.

As it is cumbersome and expensive to acquire a huge amount of data for training neural dialog models, data augmentation is proposed to effectively utilize existing training samples. However, current data augmentation techniques on the dialog generation task mostly augment all cases in the training dataset without considering the intrinsic attributes between different cases. We argue that not all cases are beneficial for augmentation task, and the cases suitable for augmentation should obey the following two attributes: (1) low-quality (the dialog model cannot generate a high-quality response for the case), (2) representative (the case should represent the property of the whole dataset). Herein, we explore this idea by proposing a Selective Data Augmentation framework (SDA) for the response generation task. SDA employs a dual adversarial network to select the lowest quality and most representative data points for augmentation in one stage. Extensive experiments conducted on two publicly available datasets, i.e., DailyDialog and OpenSubtitles, show that our framework can improve the response generation performance with respect to various metrics.

Dawei Li, Yanran Li, Jiayi Zhang et al.

Existing commonsense knowledge bases often organize tuples in an isolated manner, which is deficient for commonsense conversational models to plan the next steps. To fill the gap, we curate a large-scale multi-turn human-written conversation corpus, and create the first Chinese commonsense conversation knowledge graph which incorporates both social commonsense knowledge and dialog flow information. To show the potential of our graph, we develop a graph-conversation matching approach, and benchmark two graph-grounded conversational tasks.

Xiangru Tang, Daniel Shao, Jiwoong Sohn et al.

Large Language Models (LLMs) have shown impressive performance on existing medical question-answering benchmarks. This high performance makes it increasingly difficult to meaningfully evaluate and differentiate advanced methods. We present MedAgentsBench, a benchmark that focuses on challenging medical questions requiring multi-step clinical reasoning, diagnosis formulation, and treatment planning-scenarios where current models still struggle despite their strong performance on standard tests. Drawing from seven established medical datasets, our benchmark addresses three key limitations in existing evaluations: (1) the prevalence of straightforward questions where even base models achieve high performance, (2) inconsistent sampling and evaluation protocols across studies, and (3) lack of systematic analysis of the interplay between performance, cost, and inference time. Through experiments with various base models and reasoning methods, we demonstrate that the latest thinking models, DeepSeek R1 and OpenAI o3, exhibit exceptional performance in complex medical reasoning tasks. Additionally, advanced search-based agent methods offer promising performance-to-cost ratios compared to traditional approaches. Our analysis reveals substantial performance gaps between model families on complex questions and identifies optimal model selections for different computational constraints. Our benchmark and evaluation framework are publicly available at https://github.com/gersteinlab/medagents-benchmark.

Jinyu Xiang, Jiayi Zhang, Zhaoyang Yu et al.

Well-designed prompts are crucial for enhancing Large language models' (LLMs) reasoning capabilities while aligning their outputs with task requirements across diverse domains. However, manually designed prompts require expertise and iterative experimentation. While existing prompt optimization methods aim to automate this process, they rely heavily on external references such as ground truth or by humans, limiting their applicability in real-world scenarios where such data is unavailable or costly to obtain. To address this, we propose Self-Supervised Prompt Optimization (SPO), a cost-efficient framework that discovers effective prompts for both closed and open-ended tasks without requiring external reference. Motivated by the observations that prompt quality manifests directly in LLM outputs and LLMs can effectively assess adherence to task requirements, we derive evaluation and optimization signals purely from output comparisons. Specifically, SPO selects superior prompts through pairwise output comparisons evaluated by an LLM evaluator, followed by an LLM optimizer that aligns outputs with task requirements. Extensive experiments demonstrate that SPO outperforms state-of-the-art prompt optimization methods, achieving comparable or superior results with significantly lower costs (e.g., 1.1% to 5.6% of existing methods) and fewer samples (e.g., three samples). The code is available at https://github.com/FoundationAgents/SPO.

Jiayi Zhang, Conrad Borchers, Clayton Cohn et al.

The field of learning analytics has made notable strides in automating the detection of complex learning processes in multimodal data. However, most advancements have focused on individualized problem-solving instead of collaborative, open-ended problem-solving, which may offer both affordances (richer data) and challenges (low cohesion) to behavioral prediction. Here, we extend predictive models to automatically detect socially shared regulation of learning (SSRL) behaviors in collaborative computational modeling environments using embedding-based approaches. We leverage large language models (LLMs) as summarization tools to generate task-aware representations of student dialogue aligned with system logs. These summaries, combined with text-only embeddings, context-enriched embeddings, and log-derived features, were used to train predictive models. Results show that text-only embeddings often achieve stronger performance in detecting SSRL behaviors related to enactment or group dynamics (e.g., off-task behavior or requesting assistance). In contrast, contextual and multimodal features provide complementary benefits for constructs such as planning and reflection. Overall, our findings highlight the promise of embedding-based models for extending learning analytics by enabling scalable detection of SSRL behaviors, ultimately supporting real-time feedback and adaptive scaffolding in collaborative learning environments that teachers value.

Conrad Borchers, Jiayi Zhang, Ryan S. Baker et al.

Numerous studies demonstrate the importance of self-regulation during learning by problem-solving. Recent work in learning analytics has largely examined students' use of SRL concerning overall learning gains. Limited research has related SRL to in-the-moment performance differences among learners. The present study investigates SRL behaviors in relationship to learners' moment-by-moment performance while working with intelligent tutoring systems for stoichiometry chemistry. We demonstrate the feasibility of labeling SRL behaviors based on AI-generated think-aloud transcripts, identifying the presence or absence of four SRL categories (processing information, planning, enacting, and realizing errors) in each utterance. Using the SRL codes, we conducted regression analyses to examine how the use of SRL in terms of presence, frequency, cyclical characteristics, and recency relate to student performance on subsequent steps in multi-step problems. A model considering students' SRL cycle characteristics outperformed a model only using in-the-moment SRL assessment. In line with theoretical predictions, students' actions during earlier, process-heavy stages of SRL cycles exhibited lower moment-by-moment correctness during problem-solving than later SRL cycle stages. We discuss system re-design opportunities to add SRL support during stages of processing and paths forward for using machine learning to speed research depending on the assessment of SRL based on transcription of think-aloud data.

Jiayi Zhang, Shu Yang, Junchao Wu et al.

Fine-tuning Large Language Models on a political topic will significantly manipulate their political stance on various issues and unintentionally affect their stance on unrelated topics. While previous studies have proposed this issue, there is still a lack of understanding regarding the internal representations of these stances and the mechanisms that lead to unintended cross-topic generalization. In this paper, we systematically explore the internal mechanisms underlying this phenomenon from a neuron-level perspective and how to mitigate the cross-topic generalization of political fine-tuning. Firstly, we propose Political Neuron Localization through Activation Contrasting (PNLAC) to identify two distinct types of political neurons: general political neurons, which govern stance across multiple political topics, and topic-specific neurons} that affect the model's political stance on individual topics. We find the existence of these political neuron types across four models and datasets through activation patching experiments. Leveraging these insights, we introduce InhibitFT, an inhibition-based fine-tuning method, effectively mitigating the cross-topic stance generalization. Experimental results demonstrate the robustness of identified neuron types across various models and datasets, and show that InhibitFT significantly reduces the cross-topic stance generalization by 20% on average, while preserving topic-specific performance. Moreover, we demonstrate that selectively inhibiting only 5% of neurons is sufficient to effectively mitigate the cross-topic stance generalization.

Jiayi Zhang, Yiming Zhang, Yuan Zheng et al.

Urban traffic optimization is critical for improving transportation efficiency and alleviating congestion, particularly in large-scale dynamic networks. Traditional methods, such as Dijkstra's and Floyd's algorithms, provide effective solutions in static settings, but they struggle with the spatial-temporal complexity of real-world traffic flows. In this work, we propose TrafficKAN-GCN, a hybrid deep learning framework combining Kolmogorov-Arnold Networks (KAN) with Graph Convolutional Networks (GCN), designed to enhance urban traffic flow optimization. By integrating KAN's adaptive nonlinear function approximation with GCN's spatial graph learning capabilities, TrafficKAN-GCN captures both complex traffic patterns and topological dependencies. We evaluate the proposed framework using real-world traffic data from the Baltimore Metropolitan area. Compared with baseline models such as MLP-GCN, standard GCN, and Transformer-based approaches, TrafficKAN-GCN achieves competitive prediction accuracy while demonstrating improved robustness in handling noisy and irregular traffic data. Our experiments further highlight the framework's ability to redistribute traffic flow, mitigate congestion, and adapt to disruptive events, such as the Francis Scott Key Bridge collapse. This study contributes to the growing body of work on hybrid graph learning for intelligent transportation systems, highlighting the potential of combining KAN and GCN for real-time traffic optimization. Future work will focus on reducing computational overhead and integrating Transformer-based temporal modeling for enhanced long-term traffic prediction. The proposed TrafficKAN-GCN framework offers a promising direction for data-driven urban mobility management, balancing predictive accuracy, robustness, and computational efficiency.

Jiayi Zhang, Zhi Cui, Xiaoqiang Xia et al.

A simile is a figure of speech that directly makes a comparison, showing similarities between two different things, e.g. "Reading papers can be dull sometimes,like watching grass grow". Human writers often interpolate appropriate similes into proper locations of the plain text to vivify their writings. However, none of existing work has explored neural simile interpolation, including both locating and generation. In this paper, we propose a new task of Writing Polishment with Simile (WPS) to investigate whether machines are able to polish texts with similes as we human do. Accordingly, we design a two-staged Locate&Gen model based on transformer architecture. Our model firstly locates where the simile interpolation should happen, and then generates a location-specific simile. We also release a large-scale Chinese Simile (CS) dataset containing 5 million similes with context. The experimental results demonstrate the feasibility of WPS task and shed light on the future research directions towards better automatic text polishment.

Xiuying Chen, Zhi Cui, Jiayi Zhang et al.

In multi-turn dialog, utterances do not always take the full form of sentences \cite{Carbonell1983DiscoursePA}, which naturally makes understanding the dialog context more difficult. However, it is essential to fully grasp the dialog context to generate a reasonable response. Hence, in this paper, we propose to improve the response generation performance by examining the model's ability to answer a reading comprehension question, where the question is focused on the omitted information in the dialog. Enlightened by the multi-task learning scheme, we propose a joint framework that unifies these two tasks, sharing the same encoder to extract the common and task-invariant features with different decoders to learn task-specific features. To better fusing information from the question and the dialog history in the encoding part, we propose to augment the Transformer architecture with a memory updater, which is designed to selectively store and update the history dialog information so as to support downstream tasks. For the experiment, we employ human annotators to write and examine a large-scale dialog reading comprehension dataset. Extensive experiments are conducted on this dataset, and the results show that the proposed model brings substantial improvements over several strong baselines on both tasks. In this way, we demonstrate that reasoning can indeed help better response generation and vice versa. We release our large-scale dataset for further research.

Zhi Cui, Yanran Li, Jiayi Zhang et al.

To model diverse responses for a given post, one promising way is to introduce a latent variable into Seq2Seq models. The latent variable is supposed to capture the discourse-level information and encourage the informativeness of target responses. However, such discourse-level information is often too coarse for the decoder to be utilized. To tackle it, our idea is to transform the coarse-grained discourse-level information into fine-grained word-level information. Specifically, we firstly measure the semantic concentration of corresponding target response on the post words by introducing a fine-grained focus signal. Then, we propose a focus-constrained attention mechanism to take full advantage of focus in well aligning the input to the target response. The experimental results demonstrate that by exploiting the fine-grained signal, our model can generate more diverse and informative responses compared with several state-of-the-art models.

Jiayi Zhang, Chongyang Tao, Zhenjing Xu et al.

Generating qualitative responses has always been a challenge for human-computer dialogue systems. Existing dialogue systems generally derive from either retrieval-based or generative-based approaches, both of which have their own pros and cons. Despite the natural idea of an ensemble model of the two, existing ensemble methods only focused on leveraging one approach to enhance another, we argue however that they can be further mutually enhanced with a proper training strategy. In this paper, we propose ensembleGAN, an adversarial learning framework for enhancing a retrieval-generation ensemble model in open-domain conversation scenario. It consists of a language-model-like generator, a ranker generator, and one ranker discriminator. Aiming at generating responses that approximate the ground-truth and receive high ranking scores from the discriminator, the two generators learn to generate improved highly relevant responses and competitive unobserved candidates respectively, while the discriminative ranker is trained to identify true responses from adversarial ones, thus featuring the merits of both generator counterparts. The experimental results on a large short-text conversation data demonstrate the effectiveness of the ensembleGAN by the amelioration on both human and automatic evaluation metrics.