arXiv subjects

Chaoli Zhang

Publications and source records attributed to Chaoli Zhang.

30 records · Page 2Linked to original sources

arXiv2024

Large Language Models for Education: A Survey and Outlook

The advent of Large Language Models (LLMs) has brought in a new era of possibilities in the realm of education. This survey paper summarizes the various technologies of LLMs in educational settings from multifaceted perspectives, encompassing student and teacher assistance, adaptive learning, and commercial tools. We systematically review the technological advancements in each perspective, organize related datasets and benchmarks, and identify the risks and challenges associated with deploying LLMs in education. Furthermore, we outline future research opportunities, highlighting the potential promising directions. Our survey aims to provide a comprehensive technological picture for educators, researchers, and policymakers to harness the power of LLMs to revolutionize educational practices and foster a more effective personalized learning environment.

cs.CL↗

arXiv2023

LogiCoT: Logical Chain-of-Thought Instruction-Tuning

Generative Pre-trained Transformer 4 (GPT-4) demonstrates impressive chain-of-thought reasoning ability. Recent work on self-instruction tuning, such as Alpaca, has focused on enhancing the general proficiency of models. These instructions enable the model to achieve performance comparable to GPT-3.5 on general tasks like open-domain text generation and paraphrasing. However, they fall short of helping the model handle complex reasoning tasks. To bridge the gap, this paper presents LogiCoT, a new instruction-tuning dataset for Logical Chain-of-Thought reasoning with GPT-4. We elaborate on the process of harvesting instructions for prompting GPT-4 to generate chain-of-thought rationales. LogiCoT serves as an instruction set for teaching models of logical reasoning and elicits general reasoning skills.

cs.CL↗

arXiv2023

DCdetector: Dual Attention Contrastive Representation Learning for Time Series Anomaly Detection

Time series anomaly detection is critical for a wide range of applications. It aims to identify deviant samples from the normal sample distribution in time series. The most fundamental challenge for this task is to learn a representation map that enables effective discrimination of anomalies. Reconstruction-based methods still dominate, but the representation learning with anomalies might hurt the performance with its large abnormal loss. On the other hand, contrastive learning aims to find a representation that can clearly distinguish any instance from the others, which can bring a more natural and promising representation for time series anomaly detection. In this paper, we propose DCdetector, a multi-scale dual attention contrastive representation learning model. DCdetector utilizes a novel dual attention asymmetric design to create the permutated environment and pure contrastive loss to guide the learning process, thus learning a permutation invariant representation with superior discrimination abilities. Extensive experiments show that DCdetector achieves state-of-the-art results on multiple time series anomaly detection benchmark datasets. Code is publicly available at https://github.com/DAMO-DI-ML/KDD2023-DCdetector.

cs.LG↗

arXiv2023

Mechanism Design with Predicted Task Revenue for Bike Sharing Systems

Bike sharing systems have been widely deployed around the world in recent years. A core problem in such systems is to reposition the bikes so that the distribution of bike supply is reshaped to better match the dynamic bike demand. When the bike-sharing company or platform is able to predict the revenue of each reposition task based on historic data, an additional constraint is to cap the payment for each task below its predicted revenue. In this paper, we propose an incentive mechanism called {\em TruPreTar} to incentivize users to park bicycles at locations desired by the platform toward rebalancing supply and demand. TruPreTar possesses four important economic and computational properties such as truthfulness and budget feasibility. Furthermore, we prove that even when the payment budget is tight, the total revenue still exceeds or equals the budget. Otherwise, TruPreTar achieves 2-approximation as compared to the optimal (revenue-maximizing) solution, which is close to the lower bound of at least $\sqrt{2}$ that we also prove. Using an industrial dataset obtained from a large bike-sharing company, our experiments show that TruPreTar is effective in rebalancing bike supply and demand and, as a result, generates high revenue that outperforms several benchmark mechanisms.

cs.GT↗

arXiv2023

Transformers in Time Series: A Survey

Transformers have achieved superior performances in many tasks in natural language processing and computer vision, which also triggered great interest in the time series community. Among multiple advantages of Transformers, the ability to capture long-range dependencies and interactions is especially attractive for time series modeling, leading to exciting progress in various time series applications. In this paper, we systematically review Transformer schemes for time series modeling by highlighting their strengths as well as limitations. In particular, we examine the development of time series Transformers in two perspectives. From the perspective of network structure, we summarize the adaptations and modifications that have been made to Transformers in order to accommodate the challenges in time series analysis. From the perspective of applications, we categorize time series Transformers based on common tasks including forecasting, anomaly detection, and classification. Empirically, we perform robust analysis, model size analysis, and seasonal-trend decomposition analysis to study how Transformers perform in time series. Finally, we discuss and suggest future directions to provide useful research guidance. To the best of our knowledge, this paper is the first work to comprehensively and systematically summarize the recent advances of Transformers for modeling time series data. We hope this survey will ignite further research interests in time series Transformers.

cs.LG↗

arXiv2023

TFAD: A Decomposition Time Series Anomaly Detection Architecture with Time-Frequency Analysis

Time series anomaly detection is a challenging problem due to the complex temporal dependencies and the limited label data. Although some algorithms including both traditional and deep models have been proposed, most of them mainly focus on time-domain modeling, and do not fully utilize the information in the frequency domain of the time series data. In this paper, we propose a Time-Frequency analysis based time series Anomaly Detection model, or TFAD for short, to exploit both time and frequency domains for performance improvement. Besides, we incorporate time series decomposition and data augmentation mechanisms in the designed time-frequency architecture to further boost the abilities of performance and interpretability. Empirical studies on widely used benchmark datasets show that our approach obtains state-of-the-art performance in univariate and multivariate time series anomaly detection tasks. Code is provided at https://github.com/DAMO-DI-ML/CIKM22-TFAD.

cs.LG↗

arXiv2023

AHPA: Adaptive Horizontal Pod Autoscaling Systems on Alibaba Cloud Container Service for Kubernetes

The existing resource allocation policy for application instances in Kubernetes cannot dynamically adjust according to the requirement of business, which would cause an enormous waste of resources during fluctuations. Moreover, the emergence of new cloud services puts higher resource management requirements. This paper discusses horizontal POD resources management in Alibaba Cloud Container Services with a newly deployed AI algorithm framework named AHPA -- the adaptive horizontal pod auto-scaling system. Based on a robust decomposition forecasting algorithm and performance training model, AHPA offers an optimal pod number adjustment plan that could reduce POD resources and maintain business stability. Since being deployed in April 2021, this system has expanded to multiple customer scenarios, including logistics, social networks, AI audio and video, e-commerce, etc. Compared with the previous algorithms, AHPA solves the elastic lag problem, increasing CPU usage by 10% and reducing resource cost by more than 20%. In addition, AHPA can automatically perform flexible planning according to the predicted business volume without manual intervention, significantly saving operation and maintenance costs.

cs.LG↗

arXiv2022

NetRCA: An Effective Network Fault Cause Localization Algorithm

Localizing the root cause of network faults is crucial to network operation and maintenance. However, due to the complicated network architectures and wireless environments, as well as limited labeled data, accurately localizing the true root cause is challenging. In this paper, we propose a novel algorithm named NetRCA to deal with this problem. Firstly, we extract effective derived features from the original raw data by considering temporal, directional, attribution, and interaction characteristics. Secondly, we adopt multivariate time series similarity and label propagation to generate new training data from both labeled and unlabeled data to overcome the lack of labeled samples. Thirdly, we design an ensemble model which combines XGBoost, rule set learning, attribution model, and graph algorithm, to fully utilize all data information and enhance performance. Finally, experiments and analysis are conducted on the real-world dataset from ICASSP 2022 AIOps Challenge to demonstrate the superiority and effectiveness of our approach.

cs.LG↗

arXiv2021

Halo Mass-Concentration Relation at High-Mass End

The concentration-mass (c-M) relation encodes the key information of the assembly history of the dark matter halos, however its behavior at the high mass end has not been measured precisely in observations yet. In this paper, we report the measurement of halo c-M relation with galaxy-galaxy lensing method, using shear catalog of the Dark Energy Camera Legacy Survey (DECaLS) Data Release 8, which covers a sky area of 9500 deg^2. The foreground lenses are selected from redMaPPer, LOWZ, and CMASS catalogs, with halo mass range from 10^{13} to 10^{15} M_sun and redshift range from z=0.08 to z=0.65. We find that the concentration decreases with the halo mass from 10^{13} to 10^{14} M_sun, but shows a trend of upturn after the pivot point of ~10^{14} M_sun. We fit the measured c-M relation with the concentration model c(M)=C_0 (M/(10^{12} M_sun/h)^{-γ} [1+(M/M_0)^{0.4}], and get the values (C_0, γ, log(M_0) = (5.119_{-0.185}^{0.183}, 0.205_{-0.010}^{0.010}, 14.083_{-0.133}^{0.130}), and (4.875_{-0.208}^{0.209}, 0.221_{-0.010}^{0.010}, 13.750_{-0.141}^{0.142}) for halos with 0.08<=z<0.35 and 0.35<=z<0.65, respectively. We also show that the model including an upturn is favored over a simple power-law model. Our measurement provides important information for the recent argument of massive cluster formation process.

astro-ph.CO↗

arXiv2020

High-redshift galaxy groups as seen by Athena/WFI

The first massive galaxy groups in the Universe are predicted to have formed at redshifts well beyond two. Baryonic physics, like stellar and active galactic nuclei (AGN) feedback in this very active epoch, are expected to have left a strong imprint on the thermo-dynamic properties of these early galaxy groups. Therefore, observations of these groups are key to constrain the relative importance of these physical processes. However, current instruments are not sensitive enough to detect them easily and characterize their hot gas content. In this work, we quantify the observing power of the Advanced Telescope for High ENergy Astrophysics (Athena), the future large X-ray observatory of the European Space Agency (ESA), for discovering and characterizing early galaxy groups at high redshifts. We used the SImulation of X-ray TElescopes (SIXTE) simulator to mimic Athena observations, and a custom-made wavelet-based algorithm to detect galaxy groups and clusters in the redshift range $0.5 \le z \le 4$. We performed extensive X-ray spectral fitting in order to characterize their gas temperature and X-ray luminosity. We also investigate how well Athena will constrain different feedback mechanisms. In the deep Wide Field Imager (WFI) survey expected to be carried out during part of Athena's first four years (the nominal mission lifetime) more than 10,000 galaxy groups and clusters at $z\ge 0.5$ will be discovered. We find that Athena can detect $\sim20$ high-redshift galaxy groups with masses of $M_{500}\geq$ $5\times 10^{13}$ $M_{\odot}$ and $z\geq2$, and almost half of them will have a gas temperature determined to a precision of $ΔT/T \le 25\%$. We demonstrate that high-redshift galaxy groups can be detected very efficiently as extended sources by Athena and that a key parameter determining the total number of such newly discovered sources is the area on the sky surveyed by Athena.

astro-ph.CO↗

arXiv2016

The tight subgiant branch of the intermediate-age star cluster NGC 411 implies a single-aged stellar population

The presence of extended main-sequence turn-off (eMSTO) regions in intermediate-age star clusters in the Large and Small Magellanic Clouds is often interpreted as resulting from extended star-formation histories (SFHs), lasting $\geq$ 300 Myr. This strongly conflicts with the traditional view of the dominant star-formation mode in stellar clusters, which are thought of as single-aged stellar populations. Here we present a test of this interpretation by exploring the morphology of the subgiant branch (SGB) of NGC 411, which hosts possibly the most extended eMSTO among all known intermediate-age star clusters. We show that the width of the NGC 411 SGB favours the single-aged stellar population interpretation and rules out an extended SFH. In addition, when considering the red clump (RC) morphology and adopting the unproven premise that the widths of all features in the colour--magnitude diagram are determined by an underlying range in ages, we find that the SFH implied is still very close to that resulting from a single-aged stellar population, with a minor fraction of stars scattering to younger ages compared with the bulk of the population. The SFHs derived from the SGB and RC are both inconsistent with the SFH derived from the eMSTO region. NGC 411 has a very low escape velocity and it has unlikely undergone significant mass loss at an early stage, thus indicating that it may lack the capacity to capture most of its initial, expelled gas from stellar evolutionary processes, a condition often required for extended SFHs to take root.

astro-ph.GA↗

arXiv2015

The VMC Survey. XVIII. Radial dependence of the low-mass, 0.55--0.82 $M_\odot$ stellar mass function in the Galactic globular cluster 47 Tucanae

We use near-infrared observations obtained as part of the {\sl Visible and Infrared Survey Telescope for Astronomy} (VISTA) Survey of the Magellanic Clouds (VMC), as well as two complementary {\sl Hubble Space Telescope} ({\sl HST}) data sets, to study the luminosity and mass functions as a function of clustercentric radius of the main-sequence stars in the Galactic globular cluster 47 Tucanae. The {\sl HST} observations indicate a relative deficit in the numbers of faint stars in the central region of the cluster compared with its periphery, for $18.75\leq m_{\rm F606W}\leq 20.9$ mag (corresponding to a stellar mass range of $0.55<m_\ast/{M_\odot}<0.73$). The stellar number counts at $6.7'$ from the cluster core show a deficit for $17.62\leq m_{\rm F606W}\leq 19.7$ mag (i.e., $0.65<m_\ast/{M_\odot}<0.82$), which is consistent with expectations from mass segregation. The VMC-based stellar mass functions exhibit power-law shapes for masses in the range $0.55<m_\ast/{M_\odot}< 0.82$. These power laws are characterized by an almost constant slope, $α$. The radial distribution of the power-law slopes $α$ thus shows evidence of the importance of both mass segregation and tidal stripping, for both the first- and second-generation stars in 47 Tuc.

astro-ph.SR↗