---
license: cc-by-4.0
datasets:
- J-MADRAL/AmazonESCI
- J-MADRAL/SearchESCI
- J-MADRAL/TREC_Product_Search_2023
- J-MADRAL/TrainingData
language:
- en
metrics:
- recall
- trec_eval
base_model:
- google-bert/bert-base-uncased
pipeline_tag: text-ranking
tags:
- ProductSearch
- ReviewSearch
library_name: transformers
---


**J-BiBERT** is a BERT-sized dense retriever initialized from [BERT](https://huggingface.co/google-bert/bert-base-uncased) public checkpoint, 
further pre-trained on e-commerce product and review data, and fine-tuned jointly on product and review search retrieval tasks on the 
[Amazon ESCI](https://huggingface.co/datasets/J-MADRAL/AmazonESCI) and [Search ESCI](https://huggingface.co/datasets/J-MADRAL/SearchESCI) datasets. 
We use symmetric encoder architecture, with a single shared encoder for both queries and products.
The similarity function is *dot product*.


## Paper and Repository ##

J-BiBERT has been described in the *Multi-Aspect Joint Retrieval for E-Commerce: Bridging Product Catalogs and Customer Reviews* paper. 
The associated GitHub repository is available at [https://anonymous.4open.science/r/J-MADRAL-C4CC](https://anonymous.4open.science/r/J-MADRAL-C4CC).


## Usage (HuggingFace Transformers) ##

Using the model directly in HuggingFace transformers requires additional code available in the [repository](https://anonymous.4open.science/r/J-MADRAL-C4CC).

```python
import modeling
import torch
import transformers

# We use a couple of training queries from Amazon ESCI and Search ESCI as an example.
queries = [
    "iphone 11 pro max case"
    "cotton summer dress care instructions"
]
products_and_reviews = [
    "OtterBox Symmetry Series Case for iPhone 11 Pro Max - Black [...]",
    "Camera Lens Protector for iPhone 11 Pro/Pro Max, Tempered Glass 9H Hardness Anti-Scratch Camera Screen Protective [...]"
    "Cute, cool and comfy summer dress [...] Hand wash and line drys easily, material is crinkly so no ironing needed. [...]",
    "Excellent machine JET J-2530 15-Inch 3/4-Horsepower Bench Drill Press. Two common Amazon reviewer complaints about higher-end drill presses [...]"
]

# Load the tokenizer and model.
tokenizer = transformers.AutoTokenizer.from_pretrained("J-MADRAL/J-BiBERT")
model = modeling.BiEncoderModel.from_pretrained("J-MADRAL/J-BiBERT")

# Tokenize the input data.
q_input = tokenizer(queries,
                    add_special_tokens=True,
                    truncation=True,
                    padding=True,
                    max_length=128,
                    return_tensors="pt")
pr_input = tokenizer(products_and_reviews,
                     add_special_tokens=True,
                     truncation=True,
                     padding=True,
                     max_length=128,
                     return_tensors="pt")

# Compute embeddings: take the "pooled_output".
q_emb = model(**q_input).pooled_output
pr_emb = model(**pr_input).pooled_output

# Compute similarity scores, using dot product similarity.
scores = torch.matmul(q_emb, pr_emb.transpose(0, 1))
```


## Training Hyperparameters ##

Training Stage | Num. Epochs | Learning Rate | AP Scaling Factor | Max Num Tokens | Batch Size | Num Negatives
|---|---|---|---|---|---|---
Pre-training | 20 | 1e-4 | 0.10 | 128 | 64 | ---
Fine-tuning | 20 | 5e-6 | 0.05 | 128 | 64 | 7

The data used for fine-tuning is available at [https://huggingface.co/datasets/J-MADRAL/TrainingData](https://huggingface.co/datasets/J-MADRAL/TrainingData).


## Evaluation Results ##

#### [Amazon ESCI](https://huggingface.co/datasets/J-MADRAL/AmazonESCI) ####

Model | R@100 | R@500 | MRR | nDCG@10 | nDCG@50
|---|---|---|---|---|---
BM25 | 0.4949 | 0.6603 | 0.4056 | 0.2599 | 0.3160
[DRAGON](https://huggingface.co/facebook/dragon-plus-context-encoder) | 0.5490 | 0.7155 | 0.4541 | 0.2929 | 0.3540
[P-BiBERT](https://huggingface.co/J-MADRAL/P-BiBERT) | 0.6018 | 0.7640 | 0.4939 | 0.3276 | 0.3967
[P-MADRAL](https://huggingface.co/J-MADRAL/P-MADRAL) | **0.6235** | **0.7806** | **0.5060** | **0.3382** | **0.4104**
**J-BiBERT** | 0.5889 | 0.7556 | 0.4808 | 0.3195 | 0.3858
[J-MADRAL](https://huggingface.co/J-MADRAL/J-MADRAL) | *0.6083* | *0.7729* | *0.4988* | *0.3330* | *0.4028*

#### [TREC Product Search 2023](https://huggingface.co/datasets/J-MADRAL/TREC_Product_Search_2023) ####

Model | R@100 | R@500 | MRR | nDCG@10 | nDCG@50
|---|---|---|---|---|---
BM25 | 0.7252 | 0.8650 | 0.7746 | 0.6231 | 0.5988
[DRAGON](https://huggingface.co/facebook/dragon-plus-context-encoder) | 0.7373 | 0.8692 | 0.8135 | 0.6526 | 0.6293
[P-BiBERT](https://huggingface.co/J-MADRAL/P-BiBERT) | 0.7432 | 0.8757 | 0.8256 | 0.6612 | 0.6287
[P-MADRAL](https://huggingface.co/J-MADRAL/P-MADRAL) | **0.7547** | *0.8840* | **0.8337** | **0.6719** | **0.6480**
**J-BiBERT** | 0.7432 | 0.8684 | 0.8114 | 0.6487 | 0.6213
[J-MADRAL](https://huggingface.co/J-MADRAL/J-MADRAL) | *0.7527* | **0.8888** | *0.8333* | *0.6702* | *0.6411*

#### [Search ESCI](https://huggingface.co/datasets/J-MADRAL/SearchESCI) ####

Model | R@100 | R@500 | MRR | nDCG@10 | nDCG@50
|---|---|---|---|---|---
BM25 | 0.5875 | 0.7288 | 0.2539 | 0.2766 | 0.3101
[DRAGON](https://huggingface.co/facebook/dragon-plus-context-encoder) | 0.5451 | 0.6751 | 0.2347 | 0.2567 | 0.2873
[R-BiBERT](https://huggingface.co/J-MADRAL/R-BiBERT) | 0.5972 | 0.7350 | 0.2602 | 0.2855 | 0.3176
[R-MADRAL](https://huggingface.co/J-MADRAL/R-MADRAL) | *0.6405* | *0.7626* | *0.2944* | *0.3215* | *0.3541*
**J-BiBERT** | 0.6300 | 0.7593 | 0.2879 | 0.3140 | 0.3474
[J-MADRAL](https://huggingface.co/J-MADRAL/J-MADRAL) | **0.6488** | **0.7729** | **0.3007** | **0.3281** | **0.3611**