Update README.md

README.md

@@ -42,7 +42,7 @@ We compute for `title+" "+text` the embeddings using our `multilingual-22-12` em
 
 ## Loading the dataset
 
-In {miracl-ar-corpus-22-12](https://huggingface.co/datasets/Cohere/miracl-ar-corpus-22-12) we provide the corpus embeddings. Note, depending on the selected split, the respective files can be quite large.
+In [miracl-ar-corpus-22-12](https://huggingface.co/datasets/Cohere/miracl-ar-corpus-22-12) we provide the corpus embeddings. Note, depending on the selected split, the respective files can be quite large.
 
 You can either load the dataset like this:
 ```python
@@ -115,24 +115,38 @@ query_embedding = response.embeddings[0] # Get the embedding for the first text
 
 ## Performance
 
-In the following table we provide the nDCG@10 scores for the cohere multilingual-22-12 model in comparison to BM25 lexical search and mDPR (as provided in the [MIRACL paper](https://arxiv.org/abs/2210.09984))
-
-| Model | cohere multilingual-22-12 | BM25 lexical search | mDPR |
-|-------|---------------------------|--------------------|------|
-| miracl-ar | **64.2** | 48.1 | 49.9 |
-| miracl-bn | **61.5** | 50.8 | 44.3 |
-| miracl-es | **47.0** | 31.9 | 47.8 |
-| miracl-fa | **44.8** | 33.3 | 48 |
-| miracl-fi | **63.7** | 55.1 | 47.2 |
-| miracl-fr | **46.8** | 18.3 | 43.5 |
-| miracl-hi | **50.7** | 45.8 | 38.3 |
-| miracl-id | 44.8 | **44.9** | 27.2 |
-| miracl-ja | **49.0** | 36.9 | 43.9 |
-| miracl-ko | **50.9** | 41.9 | 41.9 |
-| miracl-ru | **49.2** | 33.4 | 40.7 |
-| miracl-sw | **61.4** | 38.3 | 29.9 |
-| miracl-te | **67.8** | 49.4 | 35.6 |
-| miracl-th | **60.2** | 48.4 | 35.8 |
-| miracl-zh | 43.8 | 18 | **51.2** |
-| **Avg** | **53.7** | 39.6 | 41.7 |
+In the following table we compare the cohere multilingual-22-12 model with Elasticsearch version 8.6.0 lexical search (title and passage indexed as independent fields). Note that Elasticsearch doesn't support all languages that are part of the MIRACL dataset.
+
+
+We compute nDCG@10 (a ranking based loss), as well as hit@3: Is at least one relevant document in the top-3 results. We find that hit@3 is easier to interpret, as it presents the number of queries for which a relevant document is found among the top-3 results.
+
+Note: MIRACL only annotated a small fraction  of passages (10 per query) for relevancy. Especially for larger Wikipedias (like English), we often found many more relevant passages. This is know as annotation holes. Real nDCG@10 and hit@3 performance is likely higher than depicted.  
+
+
+| Model | cohere multilingual-22-12 nDCG@10 | cohere multilingual-22-12 hit@3 | ES 8.6.0 nDCG@10 | ES 8.6.0 acc@3 |
+|---|---|---|---|---|
+| miracl-ar | 64.2 | 75.2 | 46.8 | 56.2 |
+| miracl-bn | 61.5 | 75.7 | 49.2 | 60.1 |
+| miracl-de | 44.4 | 60.7 | 19.6 | 29.8 |
+| miracl-en | 44.6 | 62.2 | 30.2 | 43.2 |
+| miracl-es | 47.0 | 74.1 | 27.0 | 47.2 |
+| miracl-fi | 63.7 | 76.2 | 51.4 | 61.6 |
+| miracl-fr | 46.8 | 57.1 | 17.0 | 21.6 |
+| miracl-hi | 50.7 | 62.9 | 41.0 | 48.9 |
+| miracl-id | 44.8 | 63.8 | 39.2 | 54.7 |
+| miracl-ru | 49.2 | 66.9 | 25.4 | 36.7 |
+| **Avg** | 51.7 | 67.5 | 34.7 | 46.0 |
+
+Further languages (not supported by Elasticsearch):
+| Model | cohere multilingual-22-12 nDCG@10 | cohere multilingual-22-12 hit@3 |
+|---|---|---|
+| miracl-fa | 44.8 | 53.6 |
+| miracl-ja | 49.0 | 61.0 |
+| miracl-ko | 50.9 | 64.8 |
+| miracl-sw | 61.4 | 74.5 |
+| miracl-te | 67.8 | 72.3 |
+| miracl-th | 60.2 | 71.9 |
+| miracl-yo | 56.4 | 62.2 |
+| miracl-zh | 43.8 | 56.5 |
+| **Avg** | 54.3 | 64.6 |