Causal-Driven Skill Prerequisite Structure Discovery
Authors: Shenbao Yu, Yifeng Zeng, Fan Yang, Yinghui Pan
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We demonstrate the performance of the new approach with both simulated and real-world data. The experimental results show the effectiveness of the proposed model for identifying the skills prerequisite structure. Experimental Study We evaluate the effectiveness of CSPS in recovering the prerequisite structure of skills with both simulated data (Sync1 and Sync2) and real-world data (Frc Sub and Alg0506). |
| Researcher Affiliation | Academia | Shenbao Yu1, Yifeng Zeng2*, Fan Yang1*, Yinghui Pan3* 1Department of Automation, Xiamen University, China 2Department of Computer and Information Sciences, Northumbria University, UK 3National Engineering Laboratory for Big Data System Computing Technology, Shenzhen University, China yushenbao@stu.xmu.edu.cn, yifeng.zeng@northumbria.ac.uk, yang@xmu.edu.cn, panyinghui@szu.edu.cn |
| Pseudocode | Yes | Algorithm 1: The NB-Search Algorithm, Algorithm 2: The DS-SPV Algorithm, Algorithm 3: The PMB-Search Algorithm |
| Open Source Code | No | The paper does not provide any specific links to its own source code, nor does it explicitly state that the code for their proposed method (CSPS) is available or will be released. It mentions using "semopy... Python package" and "Python package pcalg" for baselines. |
| Open Datasets | Yes | We evaluate the effectiveness of CSPS in recovering the prerequisite structure of skills with both simulated data (Sync1 and Sync2) and real-world data (Frc Sub and Alg0506). public real-world log data of the 2005-2006 curriculum Algebra I 4, abbreviated as Alg0506, which is collected from the interactions between students and computer-aided tutoring systems. 4https://pslcdatashop.web.cmu.edu/KDDCup/ |
| Dataset Splits | No | The paper describes the total sizes of its datasets (e.g., "simulated binary data with different numbers of observations (i.e., D = 150,500,1000,2000)", "535 examinees on 20 fraction-subtraction exercises measuring 8 skills", "438 students, 185 exercises, and 12 skills") but does not provide specific details on how these datasets were split into training, validation, or test sets. |
| Hardware Specification | Yes | In the following experiments, all the numerical computations are conducted on an Ubuntu server with a Core i9-1090K 3.7 GHz and 128 GB memory. |
| Software Dependencies | No | The paper mentions the use of specific Python packages for baseline implementations ("semopy" and "pcalg") but does not provide version numbers for these or any other software dependencies. For example, "we use the semopy (Igolkina and Meshcheryakov 2020) Python package to estimate all structural equation models required in the CITS, and the PC algorithm is implemented in the Python package pcalg." |
| Experiment Setup | No | The paper describes the general experimental setup, including evaluation metrics and baseline approaches, and mentions running "the average experimental results over 10 repeated trials." However, it does not provide specific hyperparameter values (e.g., learning rate, batch size, number of epochs) or detailed training configurations for the CSPS model. |