arXiv subjects

Zhipeng Chen

Publications and source records attributed to Zhipeng Chen.

At least 55 records · Page 3Linked to original sources

arXiv2024

JiuZhang3.0: Efficiently Improving Mathematical Reasoning by Training Small Data Synthesis Models

Mathematical reasoning is an important capability of large language models~(LLMs) for real-world applications. To enhance this capability, existing work either collects large-scale math-related texts for pre-training, or relies on stronger LLMs (\eg GPT-4) to synthesize massive math problems. Both types of work generally lead to large costs in training or synthesis. To reduce the cost, based on open-source available texts, we propose an efficient way that trains a small LLM for math problem synthesis, to efficiently generate sufficient high-quality pre-training data. To achieve it, we create a dataset using GPT-4 to distill its data synthesis capability into the small LLM. Concretely, we craft a set of prompts based on human education stages to guide GPT-4, to synthesize problems covering diverse math knowledge and difficulty levels. Besides, we adopt the gradient-based influence estimation method to select the most valuable math-related texts. The both are fed into GPT-4 for creating the knowledge distillation dataset to train the small LLM. We leverage it to synthesize 6 million math problems for pre-training our JiuZhang3.0 model, which only needs to invoke GPT-4 API 9.3k times and pre-train on 4.6B data. Experimental results have shown that JiuZhang3.0 achieves state-of-the-art performance on several mathematical reasoning datasets, under both natural language reasoning and tool manipulation settings. Our code and data will be publicly released in \url{https://github.com/RUCAIBox/JiuZhang3.0}.

cs.CL↗

arXiv2024

Cross-Tokamak Deployment Study of Plasma Disruption Predictors Based on Convolutional Autoencoder

In the initial stages of operation for future tokamak, facing limited data availability, deploying data-driven disruption predictors requires optimal performance with minimal use of new device data. This paper studies the issue of data utilization in data-driven disruption predictor during cross tokamak deployment. Current predictors primarily employ supervised learning methods and require a large number of disruption and non-disruption shots for training. However, the scarcity and high cost of obtaining disruption shots for future tokamaks result in imbalanced training datasets, reducing the performance of supervised learning predictors. To solve this problem, we propose the Enhanced Convolutional Autoencoder Anomaly Detection (E-CAAD) predictor. E-CAAD can be only trained by normal samples from non-disruption shots and can also be trained by disruption precursor samples when disruption shots occur. This model not only overcomes the sample imbalance in supervised learning predictors, but also overcomes the inefficient dataset utilization faced by traditional anomaly detection predictors that cannot use disruption precursor samples for training, making it more suitable for the unpredictable datasets of future tokamaks. Compared to traditional anomaly detection predictor, the E-CAAD predictor performs better in disruption prediction and is deployed faster on new devices. Additionally, we explore strategies to accelerate deployment of E-CAAD predictor on the new device by using data from existing devices. Two deployment strategies are presented: mixing data from existing devices and fine-tuning the predictor trained on existing devices. Our comparisons indicate that the data from existing device can accelerate the deployment of predictor on new device. Notably, the fine-tuning strategy yields the fastest deployment on new device among the designed strategies.

physics.plasm-ph↗

arXiv2023

Extraction of n = 0 pick-up by locked mode detectors based on neural networks in J-TEXT

Measurement of locked mode (LM) is important for the physical research of Magnetohydrodynamic (MHD) instabilities and plasma disruption. The n = 0 pick-up need to be extracted and subtracted to calculate the amplitude and phase of the LM. A new method to extract this pick-up has been developed by predicting the n = 0 pick-up brn=0 by the LM detectors based on Neural Networks (NNs) in J-TEXT. An approach called Power Multiple Time Scale (PMTS) has been developed with outstanding regressing effect in multiple frequency ranges. Three models have been progressed based on PMTS NNs. PMTS could fit the brn=0 on the LM detectors with little errors both in time domain and frequency domain. The n>0 pick-up brn>0 generated by resonant magnetic perturbations (RMPs) can be obtained after subtracting the extracted brn=0. This new method uses only one LM instead of 4 LM detectors to extract brn=0. Therefore, the distribution of the LM detectors can also be optimized based on this new method.

physics.plasm-ph↗

arXiv2023

A weighted matching scheme of magnetic coil design for FRC shaping control

The two-dimentional (2D) separatrix shaping plays a crucial role in the confinement of the Field Reversed Configuration (FRC), and the magnetic coils serve as an effective means for its control. In this work we develop a method to optimize the location, and current amplitude of the coils required for achieving the target separatrix shape. By iteratively calculating the coil currents, the plasma current, and the equilibrium magnetic flux, the equilibrium separatrix progressively converges towards the desired shape. The coil currents are determined through a matching method, and the NIMEQ code is employed to compute the FRC equilibrium with a Rigid Rotor type of plasma distribution. This approach enables the adaption of the equilibrium separatrix into any desired shape, thus offering a potential coil optimization scheme for the device design and the 2D shaping control of FRC plasma.

physics.plasm-ph↗

arXiv2023

ChatCoT: Tool-Augmented Chain-of-Thought Reasoning on Chat-based Large Language Models

Although large language models (LLMs) have achieved excellent performance in a variety of evaluation benchmarks, they still struggle in complex reasoning tasks which require specific knowledge and multi-hop reasoning. To improve the reasoning abilities, we propose ChatCoT, a tool-augmented chain-of-thought reasoning framework for chat-based LLMs (e.g., ChatGPT). In ChatCoT, we model the chain-of-thought (CoT) reasoning as multi-turn conversations, to utilize tools in a more natural way through chatting. At each turn, LLMs can either interact with tools or perform the reasoning. Our approach can effectively leverage the multi-turn conversation ability of chat-based LLMs, and integrate the thought chain following and tools manipulation in a unified way. Specially, we initialize the early turns of the conversation by the knowledge about tools, tasks, and reasoning format, and propose an iterative tool-augmented reasoning step to perform step-by-step tool-augmented reasoning. The experiment results on two complex reasoning datasets (MATH and HotpotQA) have shown the effectiveness of ChatCoT on complex reasoning tasks, achieving a 7.9% relative improvement over the state-of-the-art baseline. Our code and data are available at: \url{https://github.com/RUCAIBOX/ChatCoT}.

cs.CL↗

arXiv2023

Don't Make Your LLM an Evaluation Benchmark Cheater

Large language models~(LLMs) have greatly advanced the frontiers of artificial intelligence, attaining remarkable improvement in model capacity. To assess the model performance, a typical approach is to construct evaluation benchmarks for measuring the ability level of LLMs in different aspects. Despite that a number of high-quality benchmarks have been released, the concerns about the appropriate use of these benchmarks and the fair comparison of different models are increasingly growing. Considering these concerns, in this paper, we discuss the potential risk and impact of inappropriately using evaluation benchmarks and misleadingly interpreting the evaluation results. Specially, we focus on a special issue that would lead to inappropriate evaluation, \ie \emph{benchmark leakage}, referring that the data related to evaluation sets is occasionally used for model training. This phenomenon now becomes more common since pre-training data is often prepared ahead of model test. We conduct extensive experiments to study the effect of benchmark leverage, and find that it can dramatically boost the evaluation results, which would finally lead to an unreliable assessment of model performance. To improve the use of existing evaluation benchmarks, we finally present several guidelines for both LLM developers and benchmark maintainers. We hope this work can draw attention to appropriate training and evaluation of LLMs.

cs.CL↗

arXiv2023

Electrical Characteristics of the GEC Reference Cell with Impedance Matching: A Two-Dimensional PIC/MCC Modeling Study

In this paper, the electrical characteristics of the Gaseous Electronics Conference (GEC) reference cell with impedance matching are investigated through a two-dimensional electrostatic implicit Particle-in-Cell/Monte Carlo Collision (PIC/MCC) model in an axisymmetric coordinate system. The coupling between the complex reactor geometry and the external circuit is included via an equivalent capacitance calculated from the electric energy density. The results of this model are compared with experimental measurements and other model calculations and show good agreement. This simulation obtains the plasma kinetics of the capacitively coupled discharge process at low pressure and detailed external circuit responses, including power transmission, reflection, and higher-order harmonics in the circuit, which provides important insights for impedance-matching design in semiconductor plasma processing.

physics.plasm-ph↗

arXiv2023

Critical roles of edge turbulent transport in the formation of high-field-side high-density front and density limit disruption in J-TEXT tokamak

This article presents an in-depth study of the sequence of events leading to density limit disruption in J-TEXT tokamak plasmas, with an emphasis on boudary turbulent transport and the high-field-side high-density (HFSHD) front. These phenomena were extensively investigated by using Langmuir probe and Polarimeter-interferometer diagnostics.

physics.plasm-ph↗

arXiv2023

JiuZhang 2.0: A Unified Chinese Pre-trained Language Model for Multi-task Mathematical Problem Solving

Although pre-trained language models~(PLMs) have recently advanced the research progress in mathematical reasoning, they are not specially designed as a capable multi-task solver, suffering from high cost for multi-task deployment (\eg a model copy for a task) and inferior performance on complex mathematical problems in practical applications. To address these issues, in this paper, we propose \textbf{JiuZhang~2.0}, a unified Chinese PLM specially for multi-task mathematical problem solving. Our idea is to maintain a moderate-sized model and employ the \emph{cross-task knowledge sharing} to improve the model capacity in a multi-task setting. Specially, we construct a Mixture-of-Experts~(MoE) architecture for modeling mathematical text, so as to capture the common mathematical knowledge across tasks. For optimizing the MoE architecture, we design \emph{multi-task continual pre-training} and \emph{multi-task fine-tuning} strategies for multi-task adaptation. These training strategies can effectively decompose the knowledge from the task data and establish the cross-task sharing via expert networks. In order to further improve the general capacity of solving different complex tasks, we leverage large language models~(LLMs) as complementary models to iteratively refine the generated solution by our PLM, via in-context learning. Extensive experiments have demonstrated the effectiveness of our model.

cs.CL↗

arXiv2023

Validation of the plasma-wall self-organization model for density limit in ECRH-assisted start-up of Ohmic discharges on J-TEXT

A recently developed plasma-wall self-organization (PWSO) model predicts a significantly enhanced density limit, which may be attainable in tokamaks with ECRH-assisted ohmic startup and sufficiently high initial neutral density. Experiments have been conducted on J-TEXT to validate such a density limit scenario based on this model. Experimental results demonstrate that increasing the pre-filled gas pressure or ECRH power during the startup phase can effectively enhance plasma purity and raise the density limit at the flat-top. Despite the dominant carbon fraction in the wall material, some discharges approach the edge of the density-free regime of the 1D model of PWSO.

physics.plasm-ph↗

arXiv2023

Disruption Precursor Onset Time Study Based on Semi-supervised Anomaly Detection

The full understanding of plasma disruption in tokamaks is currently lacking, and data-driven methods are extensively used for disruption prediction. However, most existing data-driven disruption predictors employ supervised learning techniques, which require labeled training data. The manual labeling of disruption precursors is a tedious and challenging task, as some precursors are difficult to accurately identify, limiting the potential of machine learning models. To address this issue, commonly used labeling methods assume that the precursor onset occurs at a fixed time before the disruption, which may not be consistent for different types of disruptions or even the same type of disruption, due to the different speeds at which plasma instabilities escalate. This leads to mislabeled samples and suboptimal performance of the supervised learning predictor. In this paper, we present a disruption prediction method based on anomaly detection that overcomes the drawbacks of unbalanced positive and negative data samples and inaccurately labeled disruption precursor samples. We demonstrate the effectiveness and reliability of anomaly detection predictors based on different algorithms on J-TEXT and EAST to evaluate the reliability of the precursor onset time inferred by the anomaly detection predictor. The precursor onset times inferred by these predictors reveal that the labeling methods have room for improvement as the onset times of different shots are not necessarily the same. Finally, we optimize precursor labeling using the onset times inferred by the anomaly detection predictor and test the optimized labels on supervised learning disruption predictors. The results on J-TEXT and EAST show that the models trained on the optimized labels outperform those trained on fixed onset time labels.

physics.plasm-ph↗

arXiv2022

TextBox 2.0: A Text Generation Library with Pre-trained Language Models

To facilitate research on text generation, this paper presents a comprehensive and unified library, TextBox 2.0, focusing on the use of pre-trained language models (PLMs). To be comprehensive, our library covers $13$ common text generation tasks and their corresponding $83$ datasets and further incorporates $45$ PLMs covering general, translation, Chinese, dialogue, controllable, distilled, prompting, and lightweight PLMs. We also implement $4$ efficient training strategies and provide $4$ generation objectives for pre-training new PLMs from scratch. To be unified, we design the interfaces to support the entire research pipeline (from data loading to training and evaluation), ensuring that each step can be fulfilled in a unified way. Despite the rich functionality, it is easy to use our library, either through the friendly Python API or command line. To validate the effectiveness of our library, we conduct extensive experiments and exemplify four types of research scenarios. The project is released at the link: https://github.com/RUCAIBox/TextBox.

cs.CL↗

arXiv2022

Transferable Cross-Tokamak Disruption Prediction with Deep Hybrid Neural Network Feature Extractor

Predicting disruptions across different tokamaks is a great obstacle to overcome. Future tokamaks can hardly tolerate disruptions at high performance discharge. Few disruption discharges at high performance can hardly compose an abundant training set, which makes it difficult for current data-driven methods to obtain an acceptable result. A machine learning method capable of transferring a disruption prediction model trained on one tokamak to another is required to solve the problem. The key is a disruption prediction model containing a feature extractor that is able to extract common disruption precursor traces in tokamak diagnostic data, and a transferable disruption classifier. Based on the concerns above, the paper first presents a deep fusion feature extractor designed specifically for extracting disruption precursor features from common diagnostics on tokamaks according to currently known precursors of disruption, providing a promising foundation for transferable models. The fusion feature extractor is proved by comparing with manual feature extraction on J-TEXT. Based on the feature extractor trained on J-TEXT, the disruption prediction model was transferred to EAST data with mere 20 discharges from EAST experiment. The performance is comparable with a model trained with 1896 discharges from EAST. From the comparison among other model training scenarios, transfer learning showed its potential in predicting disruptions across different tokamaks.

physics.plasm-ph↗

arXiv2022

ElitePLM: An Empirical Study on General Language Ability Evaluation of Pretrained Language Models

Nowadays, pretrained language models (PLMs) have dominated the majority of NLP tasks. While, little research has been conducted on systematically evaluating the language abilities of PLMs. In this paper, we present a large-scale empirical study on general language ability evaluation of PLMs (ElitePLM). In our study, we design four evaluation dimensions, i.e. memory, comprehension, reasoning, and composition, to measure ten widely-used PLMs within five categories. Our empirical results demonstrate that: (1) PLMs with varying training objectives and strategies are good at different ability tests; (2) fine-tuning PLMs in downstream tasks is usually sensitive to the data size and distribution; (3) PLMs have excellent transferability between similar tasks. Moreover, the prediction results of PLMs in our experiments are released as an open resource for more deep and detailed analysis on the language abilities of PLMs. This paper can guide the future work to select, apply, and design PLMs for specific tasks. We have made all the details of experiments publicly available at https://github.com/RUCAIBox/ElitePLM.

cs.CL↗

arXiv2021

MHD analysis on the physical designs of CFETR and HFRC

The China Fusion Engineering Test Reactor (CFETR) and the Huazhong Field Reversed Configuration (HFRC), currently both under intensive physical and engineering designs in China, are the two major projects representative of the low-density steady-state and high-density pulsed pathways to fusion. One of the primary tasks of the physics designs for both CFETR and HFRC is the assessment and analysis of the magnetohydrodynamic (MHD) stability of the proposed design schemes. Comprehensive efforts on the assessment of MHD stability of CFETR and HFRC baseline scenarios have led to preliminary progresses that may further benefit engineering designs.

physics.plasm-ph↗

arXiv2021

TextBox: A Unified, Modularized, and Extensible Framework for Text Generation

In this paper, we release an open-source library, called TextBox, to provide a unified, modularized, and extensible text generation framework. TextBox aims to support a broad set of text generation tasks and models. In our library, we implement 21 text generation models on 9 benchmark datasets, covering the categories of VAE, GAN, and pretrained language models. Meanwhile, our library maintains sufficient modularity and extensibility by properly decomposing the model architecture, inference, and learning process into highly reusable modules, which allows users to easily incorporate new models into our framework. The above features make TextBox specially suitable for researchers and practitioners to quickly reproduce baseline models and develop new models. TextBox is implemented based on PyTorch, and released under Apache License 2.0 at https://github.com/RUCAIBox/TextBox.

cs.AI↗

arXiv2020

A Sentence Cloze Dataset for Chinese Machine Reading Comprehension

Owing to the continuous efforts by the Chinese NLP community, more and more Chinese machine reading comprehension datasets become available. To add diversity in this area, in this paper, we propose a new task called Sentence Cloze-style Machine Reading Comprehension (SC-MRC). The proposed task aims to fill the right candidate sentence into the passage that has several blanks. We built a Chinese dataset called CMRC 2019 to evaluate the difficulty of the SC-MRC task. Moreover, to add more difficulties, we also made fake candidates that are similar to the correct ones, which requires the machine to judge their correctness in the context. The proposed dataset contains over 100K blanks (questions) within over 10K passages, which was originated from Chinese narrative stories. To evaluate the dataset, we implement several baseline systems based on the pre-trained models, and the results show that the state-of-the-art model still underperforms human performance by a large margin. We release the dataset and baseline system to further facilitate our community. Resources available through https://github.com/ymcui/cmrc2019

cs.CL↗

arXiv2020

TextBrewer: An Open-Source Knowledge Distillation Toolkit for Natural Language Processing

In this paper, we introduce TextBrewer, an open-source knowledge distillation toolkit designed for natural language processing. It works with different neural network models and supports various kinds of supervised learning tasks, such as text classification, reading comprehension, sequence labeling. TextBrewer provides a simple and uniform workflow that enables quick setting up of distillation experiments with highly flexible configurations. It offers a set of predefined distillation methods and can be extended with custom code. As a case study, we use TextBrewer to distill BERT on several typical NLP tasks. With simple configurations, we achieve results that are comparable with or even higher than the public distilled BERT models with similar numbers of parameters. Our toolkit is available through: http://textbrewer.hfl-rc.com

cs.CL↗