Yi Hu

Jiabang He, Lei Wang, Yi Hu et al.

Large language models (LLMs), such as GPT-3 and ChatGPT, have demonstrated remarkable results in various natural language processing (NLP) tasks with in-context learning, which involves inference based on a few demonstration examples. Despite their successes in NLP tasks, no investigation has been conducted to assess the ability of LLMs to perform document information extraction (DIE) using in-context learning. Applying LLMs to DIE poses two challenges: the modality and task gap. To this end, we propose a simple but effective in-context learning framework called ICL-D3IE, which enables LLMs to perform DIE with different types of demonstration examples. Specifically, we extract the most difficult and distinct segments from hard training documents as hard demonstrations for benefiting all test instances. We design demonstrations describing relationships that enable LLMs to understand positional relationships. We introduce formatting demonstrations for easy answer extraction. Additionally, the framework improves diverse demonstrations by updating them iteratively. Our experiments on three widely used benchmark datasets demonstrate that the ICL-D3IE framework enables Davinci-003/ChatGPT to achieve superior performance when compared to previous pre-trained methods fine-tuned with full training in both the in-distribution (ID) setting and in the out-of-distribution (OOD) setting. Code is available at https://github.com/MAEHCM/ICL-D3IE.

Jiabang He, Yi Hu, Lei Wang et al.

Numerous pre-training techniques for visual document understanding (VDU) have recently shown substantial improvements in performance across a wide range of document tasks. However, these pre-trained VDU models cannot guarantee continued success when the distribution of test data differs from the distribution of training data. In this paper, to investigate how robust existing pre-trained VDU models are to various distribution shifts, we first develop an out-of-distribution (OOD) benchmark termed Do-GOOD for the fine-Grained analysis on Document image-related tasks specifically. The Do-GOOD benchmark defines the underlying mechanisms that result in different distribution shifts and contains 9 OOD datasets covering 3 VDU related tasks, e.g., document information extraction, classification and question answering. We then evaluate the robustness and perform a fine-grained analysis of 5 latest VDU pre-trained models and 2 typical OOD generalization algorithms on these OOD datasets. Results from the experiments demonstrate that there is a significant performance gap between the in-distribution (ID) and OOD settings for document images, and that fine-grained analysis of distribution shifts can reveal the brittle nature of existing pre-trained VDU models and OOD generalization algorithms. The code and datasets for our Do-GOOD benchmark can be found at https://github.com/MAEHCM/Do-GOOD.

Yi Hu, Jinhang Zuo, Bob Iannucci et al. · cmu

Internet of Things (IoT) technologies have enabled numerous data-driven mobile applications and have the potential to significantly improve environmental monitoring and hazard warnings through the deployment of a network of IoT sensors. However, these IoT devices are often power-constrained and utilize wireless communication schemes with limited bandwidth. Such power constraints limit the amount of information each device can share across the network, while bandwidth limitations hinder sensors' coordination of their transmissions. In this work, we formulate the communication planning problem of IoT sensors that track the state of the environment. We seek to optimize sensors' decisions in collecting environmental data under stringent resource constraints. We propose a multi-agent reinforcement learning (MARL) method to find the optimal communication policies for each sensor that maximize the tracking accuracy subject to the power and bandwidth limitations. MARL learns and exploits the spatial-temporal correlation of the environmental data at each sensor's location to reduce the redundant reports from the sensors. Experiments on wildfire spread with LoRA wireless network simulators show that our MARL method can learn to balance the need to collect enough data to predict wildfire spread with unknown bandwidth limitations.

Yi Hu, Bo Cai, Yaoxiang Yu

Despite the continuous efforts in improving both the effectiveness and efficiency of code search, two issues remained unsolved. First, programming languages have inherent strong structural linkages, and feature mining of code as text form would omit the structural information contained inside it. Second, there is a potential semantic relationship between code and query, it is challenging to align code and text across sequences so that vectors are spatially consistent during similarity matching. To tackle both issues, in this paper, a code search model named CSSAM (Code Semantics and Structures Attention Matching) is proposed. By introducing semantic and structural matching mechanisms, CSSAM effectively extracts and fuses multidimensional code features. Specifically, the cross and residual layer was developed to facilitate high-latitude spatial alignment of code and query at the token level. By leveraging the residual interaction, a matching module is designed to preserve more code semantics and descriptive features, that enhances the adhesion between the code and its corresponding query text. Besides, to improve the model's comprehension of the code's inherent structure, a code representation structure named CSRG (Code Semantic Representation Graph) is proposed for jointly representing abstract syntax tree nodes and the data flow of the codes. According to the experimental results on two publicly available datasets containing 540k and 330k code segments, CSSAM significantly outperforms the baselines in terms of achieving the highest SR@1/5/10, MRR, and NDCG@50 on both datasets respectively. Moreover, the ablation study is conducted to quantitatively measure the impact of each key component of CSSAM on the efficiency and effectiveness of code search, which offers the insights into the improvement of advanced code search solutions.

Yi Hu, Cai Zhou, Muhan Zhang

The scaling of large language models (LLMs) emphasizes increasing depth, yet performance gains diminish with added layers. Prior work introduces the concept of "effective depth", arguing that deeper models fail to fully utilize their layers for meaningful computation. Building on this, we systematically study how effective depth varies with model scale, training type, and task difficulty. First, we analyze the model behavior of Qwen-2.5 family (1.5B-32B) and find that while the number of effective layers grows with model size, the effective depth ratio remains stable. Besides, comparisons between base and corresponding long-CoT models show no increase in effective depth, suggesting that improved reasoning stems from longer context rather than deeper per-token computation. Furthermore, evaluations across tasks of varying difficulty indicate that models do not dynamically use more layers for harder problems. Our results suggest that current LLMs underuse available depth across scales, training paradigms and tasks of varying difficulties, pointing out research opportunities on increasing the layer utilization rate of LLMs, model pruning, and early exiting. Our code is released at https://github.com/AheadOFpotato/what_affects_effective_depth.

Haotong Yang, Yi Hu, Shijia Kang et al.

Large language models (LLMs) can solve an increasing number of complex reasoning tasks while making surprising mistakes in basic numerical understanding and processing (such as 9.11 > 9.9). The latter ability is essential for tackling complex arithmetic and mathematical problems and serves as a foundation for most reasoning tasks, but previous work paid little attention to it or only discussed several restricted tasks (like integer addition). In this paper, we comprehensively investigate the numerical understanding and processing ability (NUPA) of LLMs. Firstly, we introduce a benchmark covering four common numerical representations and 17 distinct numerical tasks in four major categories, resulting in 41 meaningful combinations in total. These tasks are derived from primary and secondary education curricula, encompassing nearly all everyday numerical understanding and processing scenarios, and the rules of these tasks are very simple and clear. Through the benchmark, we find that current LLMs fail frequently in many of the tasks. To study the problem, we train small models with existing and potential techniques for enhancing NUPA (such as tokenizers, PEs, and number formats), comprehensively evaluating their effectiveness using our testbed. We also finetune practical-scale LLMs on our proposed NUPA tasks and find that 1) naive finetuning can improve NUPA a lot on many but not all tasks, and 2) surprisingly, techniques designed to enhance NUPA prove ineffective for finetuning pretrained models. We further explore the impact of chain-of-thought techniques on NUPA. Our work provides a more detailed and comprehensive understanding of NUPA in LLMs. Our benchmark and code are released at https://github.com/GraphPKU/number_cookbook.

Ningyu Zhang, Bozhong Tian, Siyuan Cheng et al.

Knowledge editing for large language models can offer an efficient solution to alter a model's behavior without negatively impacting the overall performance. However, the current approaches encounter issues with limited generalizability across tasks, necessitating one distinct editor for each task, significantly hindering the broader applications. To address this, we take the first step to analyze the multi-task generalization issue in knowledge editing. Specifically, we develop an instruction-based editing technique, termed InstructEdit, which facilitates the editor's adaptation to various task performances simultaneously using simple instructions. With only one unified editor for each LLM, we empirically demonstrate that InstructEdit can improve the editor's control, leading to an average 14.86% increase in Reliability in multi-task editing setting. Furthermore, experiments involving holdout unseen task illustrate that InstructEdit consistently surpass previous strong baselines. To further investigate the underlying mechanisms of instruction-based knowledge editing, we analyze the principal components of the editing gradient directions, which unveils that instructions can help control optimization direction with stronger OOD generalization. Code and datasets are available in https://github.com/zjunlp/EasyEdit.

Yi Hu, Shijia Kang, Haotong Yang et al.

Length generalization, the ability to solve problems longer than those seen during training, remains a critical challenge for large language models (LLMs). Previous work modifies positional encodings (PEs) and data formats to improve length generalization on specific symbolic tasks such as addition and sorting. However, these approaches are fundamentally limited to special tasks, often degrading general language performance. Furthermore, they are typically evaluated on small transformers trained from scratch on single tasks and can cause performance drop when applied during post-training stage of practical LLMs with general capabilities. Hu et al., (2024) proposed Rule-Following Fine-Tuning (RFFT) to improve length generalization in the post-training stage of LLMs. Despite its compatibility with practical models and strong performance, RFFT is proposed for single tasks too, requiring re-training for each individual task with extensive examples. In this paper, we study length generalization in multi-task settings and propose Meta Rule-Following Fine-Tuning (Meta-RFFT), the first framework enabling robust cross-task length generalization. As our first contribution, we construct a large length generalization dataset containing 86 tasks spanning code execution, number processing, symbolic and logical reasoning tasks, beyond the common addition or multiplication tasks. Secondly, we show that cross-task length generalization is possible with Meta-RFFT. After training on a large number of tasks and instances, the models achieve remarkable length generalization ability on unseen tasks with minimal fine-tuning or one-shot prompting. For example, after fine-tuning on 1 to 5 digit addition, our 32B model achieves 95% accuracy on 30 digit addition, significantly outperforming the state-of-the-art reasoning models (DeepSeek-R1-671B: 72%), despite never seeing this task during RF-pretraining.

Yi Hu, Zheyuan Cheng

Reliable detection and classification of power system events are critical for maintaining grid stability and situational awareness. Existing approaches often depend on limited labeled datasets, which restricts their ability to generalize to rare or unseen disturbances. This paper proposes a novel framework that integrates generative modeling, sliding-window temporal processing, and decision fusion to achieve robust event detection and classification using synchrophasor data. A variational autoencoder-generative adversarial network is employed to model normal operating conditions, where both reconstruction error and discriminator error are extracted as anomaly indicators. Two complementary decision strategies are developed: a threshold-based rule for computational efficiency and a convex hull-based method for robustness under complex error distributions. These features are organized into spatiotemporal detection and classification matrices through a sliding-window mechanism, and an identification and decision fusion stage integrates the outputs across PMUs. This design enables the framework to identify known events while systematically classifying previously unseen disturbances into a new category, addressing a key limitation of supervised classifiers. Experimental results demonstrate state-of-the-art accuracy, surpassing machine learning, deep learning, and envelope-based baselines. The ability to recognize unknown events further highlights the adaptability and practical value of the proposed approach for wide-area event analysis in modern power systems.

Yi Hu, Hanchi Ren, Jingjing Deng et al.

Stock price prediction is a critical area of financial forecasting, traditionally approached by training models using the historical price data of individual stocks. While these models effectively capture single-stock patterns, they fail to leverage potential correlations among stock trends, which could improve predictive performance. Current single-stock learning methods are thus limited in their ability to provide a broader understanding of price dynamics across multiple stocks. To address this, we propose a novel method that merges local patterns into a global understanding through cross-stock pattern integration. Our strategy is inspired by Federated Learning (FL), a paradigm designed for decentralized model training. FL enables collaborative learning across distributed datasets without sharing raw data, facilitating the aggregation of global insights while preserving data privacy. In our adaptation, we train models on individual stock data and iteratively merge them to create a unified global model. This global model is subsequently fine-tuned on specific stock data to retain local relevance. The proposed strategy enables parallel training of individual stock models, facilitating efficient utilization of computational resources and reducing overall training time. We conducted extensive experiments to evaluate the proposed method, demonstrating that it outperforms benchmark models and enhances the predictive capabilities of state-of-the-art approaches. Our results highlight the efficacy of Cross-Stock Trend Integration (CSTI) in advancing stock price prediction, offering a robust alternative to traditional single-stock learning methodologies.

Yi Hu, Haotong Yang, Zhouchen Lin et al.

Large language models (LLMs) have scaled up to unlock a wide range of complex reasoning tasks with the aid of various prompting methods. However, current prompting methods generate natural language intermediate steps to help reasoning, which can cause imperfect task reduction and confusion. To mitigate such limitations, we explore code prompting, a neural symbolic prompting method with both zero-shot and few-shot versions which triggers code as intermediate steps. We conduct experiments on 7 widely-used benchmarks involving symbolic reasoning and arithmetic reasoning. Code prompting generally outperforms chain-of-thought (CoT) prompting. To further understand the performance and limitations of code prompting, we perform extensive ablation studies and error analyses, and identify several exclusive advantages of using symbolic promptings compared to natural language. We also consider the ensemble of code prompting and CoT prompting to combine the strengths of both. Finally, we show through experiments how code annotations and their locations affect code prompting.

Yi Hu, Chaoran Zhang, Edward Andert et al.

Careful placement of a computational application within a target device cluster is critical for achieving low application completion time. The problem is challenging due to its NP-hardness and combinatorial nature. In recent years, learning-based approaches have been proposed to learn a placement policy that can be applied to unseen applications, motivated by the problem of placing a neural network across cloud servers. These approaches, however, generally assume the device cluster is fixed, which is not the case in mobile or edge computing settings, where heterogeneous devices move in and out of range for a particular application. We propose a new learning approach called GiPH, which learns policies that generalize to dynamic device clusters via 1) a novel graph representation gpNet that efficiently encodes the information needed for choosing a good placement, and 2) a scalable graph neural network (GNN) that learns a summary of the gpNet information. GiPH turns the placement problem into that of finding a sequence of placement improvements, learning a policy for selecting this sequence that scales to problems of arbitrary size. We evaluate GiPH with a wide range of task graphs and device clusters and show that our learned policy rapidly find good placements for new problem instances. GiPH finds placements with up to 30.5% lower completion times, searching up to 3X faster than other search-based placement policies.