Update README.md

README.md

@@ -1,16 +1,36 @@
-## Step 1: Prepare the datasets
+# Softmax-CPR <img src="https://github.com/iesl/Softmax-CPR/blob/main/imgs/automated-external-defibrillators-g7991e1588_640.png?raw=true" width="20" height="20"> for Sequential Recommendation
 
-Download the datasets from https://drive.google.com/drive/folders/1ahiLmzU7cGRPXf5qGMqtAChte2eYp9gI and put the data .inter files into the corresponding folders in ./dataset
+<p align="center"><img src="https://github.com/iesl/softmax_CPR_recommend/blob/master/softmax_limits.png?raw=true" width="651" height="365"></p>
+
+## Introduction
+
+Recent studies suggest that the existing neural models have difficulty handling repeated items in sequential recommendation tasks. However, our understanding of this difficulty is still limited. In this study, we substantially advance this field by identifying a major source of the problem: the single hidden state embedding and static item embeddings in the output softmax layer. Specifically, the similarity structure of the global item embedding in the softmax layer sometimes forces the single hidden state embedding to be close to new items when copying is a better choice, while sometimes forcing the hidden state to be close to the items from the input inappropriately. To alleviate the problem, we adapt the recently-proposed softmax alternatives such as softmax-CPR to sequential recommendation tasks and demonstrate that the new softmax architectures unleash the capability of the neural encoder on learning when to copy and when to exclude the items from the input sequence. By only making some simple modifications on the output softmax layer for SASRec and GRU4Rec, softmax-CPR achieves consistent improvement in 12 datasets. With almost the same model size, our best method not only improves the average NDCG@10 of GRU4Rec in 5 datasets with duplicated items by 10% (4%-17% individually) but also improves 7 datasets without duplicated items by 24% (8%-39%)!
+
+## How to Run
+
+### Step 1: Prepare the datasets
+
+Download the datasets from RecBole ([Google Drive](https://drive.google.com/drive/folders/1ahiLmzU7cGRPXf5qGMqtAChte2eYp9gI) or https://github.com/RUCAIBox/RecSysDatasets) and put the data .inter files into the corresponding folders in ./dataset
 
 Our code expect the folder names in ./dataset to be algebra2008_2009, amazon-beauty, Amazon_Books, Amazon_Video_Games, gowalla, ml-10m,  ml-1m, steam, tmall-buy, Twitch-100k, yelp2018, and yoochoose-clicks
 
 Each folder should contain one file called {folder_name}.inter (e.g., amazon-beauty.inter) 
 
-## Step 2: Run the code
+### Step 2: Run the code
 
 If your server supports slurm, you can just run run_hyper_slurm.sh (parallelly). Otherwise, run run_hyper_loop.sh (sequentially)
 
-You can change the variables (param_arr, model, and model_name) in the script to switch between GRU4Rec and SASRec, and change the config (model_config and model_config_name) to switch the softmax alternatives
+You can change the variables (model_name) in the script to switch between GRU4Rec, SASRec, and RepeatNet, and change the config (softmax_mode) to switch the softmax alternatives.
 
-## RecoBole Support
+### RecoBole Support
 If you encounter questions about Recbole, check the readme of recbole: README_RECBOLE_EN.md (English) or README_RECBOLE_CN.md (Chinese)
+
+Notice that our current code does not support the negative sampling during the training or inference time.
+
+## Questions
+If you have any questions or find any bugs, please send an email to Haw-Shiuan Chang (ken77921@gmail.com).
+
+## Citation
+If you use the code or Softmax-CPR in your work, please cite the following paper:
+
+Haw-Shiuan Chang, Nikhil Agarwal, and Andrew McCallum. 2024. To Copy, or not to Copy; That is a Critical Issue of the Output Softmax Layer in Neural Sequential Recommenders. In Proceedings of The 17th ACM Inernational Conference on Web Search and Data Mining (WSDM 24).