Optimize Incompatible Parameters Through Compatibility-aware Knowledge Integration

Authors: Zheqi Lv, Keming Ye, Zishu Wei, Qi Tian, Shengyu Zhang, Wenqiao Zhang, Wenjie Wang, Kun Kuang, Tat-Seng Chua, Fei Wu

AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive experiments on various recommendation and language datasets show that CKI can effectively optimize incompatible parameters under multiple tasks and settings to break through the training limit of the original model without increasing the inference cost.
Researcher Affiliation Collaboration 1Zhejiang University, Hangzhou, China 2Tencent TEG., Shenzhen, China 3 National University of Singapore, Singapore EMAIL, EMAIL, EMAIL, EMAIL
Pseudocode No The paper includes a methodological section with descriptions and a diagram (Figure 2) outlining the CKI process, but it does not present any formal pseudocode or algorithm blocks.
Open Source Code No The paper does not contain any explicit statement about releasing source code, a link to a code repository, or information that the code is provided in supplementary materials.
Open Datasets Yes In recommendation tasks, we evaluated our method on 4 widely used datasets Amazon-Beauty (Beauty), Douban-Book (Book), Douban-Music (Music), Movielens-1M (Movielens). In language tasks, we evaluated our method on 2 widely used datasets RTE, and SST-2.
Dataset Splits No The paper mentions the datasets used but does not specify the exact training, validation, or test splits (e.g., percentages, sample counts, or references to predefined splits with citations) needed for reproducibility.
Hardware Specification No The paper does not provide any specific details about the hardware used to conduct the experiments, such as GPU models, CPU types, or memory.
Software Dependencies No The paper mentions various models and methods but does not specify the software dependencies (e.g., programming languages, libraries, frameworks) along with their version numbers that were used for implementation or experimentation.
Experiment Setup No The paper discusses the experimental results and states that the integrated model needs "only one epoch of retraining to achieve better performance" for fine-tuning. However, it does not provide specific hyperparameters like learning rates, batch sizes, optimizers, or other detailed training configurations for the initial model training or fine-tuning.