hwaaaaanii commited on
Commit
d40f339
·
verified ·
1 Parent(s): 45d0ca5

Note the KDD 2027 acceptance and clarify the sources field

Browse files
Files changed (1) hide show
  1. README.md +4 -2
README.md CHANGED
@@ -57,6 +57,8 @@ configs:
57
 
58
  # RetailBridge
59
 
 
 
60
  **RetailBridge** is a multi-source question-answering benchmark for enterprise retail.
61
  Each question is answered over a heterogeneous retail database with five SQL tables and
62
  three free-text stores (reviews, customer inquiries, live-commerce sales scripts).
@@ -149,7 +151,7 @@ questions through two operators.
149
  | `answer` | string | Gold answer |
150
  | `reasoning_level` | int | 1, 2 or 3 |
151
  | `source_chaining` | string | `Homo` or `Hetero` |
152
- | `sources` | list[string] | Gold evidence sources in hop order, used as the reference for source-routing F1 (see *Source names*) |
153
  | `evidence_modality` | string | `sql-only`, `text-only` or `composite` |
154
  | `composite_family` | string or null | `entity_linked` or `answer_linked`; null for Level 1 |
155
  | `operation_type` | string | Operation of the final hop (`search`, `aggregate`, `filter`, `rank`, `calculate`, `compare`) |
@@ -218,7 +220,7 @@ the SQL tables.
218
 
219
  The [code repository](https://github.com/DISL-Lab/RetailBridge) contains RetailRAG and
220
  the evaluation script. Answer accuracy is judged by an LLM (Muse Ember) against the gold
221
- answer; source-routing F1 compares the sources an agent queries with `sources`; results for 16 LLM backbones are on the
222
  [project page](https://retailbridge-kdd.github.io).
223
 
224
  ## License
 
57
 
58
  # RetailBridge
59
 
60
+ **Accepted at KDD 2027 (Datasets and Benchmarks Track).**
61
+
62
  **RetailBridge** is a multi-source question-answering benchmark for enterprise retail.
63
  Each question is answered over a heterogeneous retail database with five SQL tables and
64
  three free-text stores (reviews, customer inquiries, live-commerce sales scripts).
 
151
  | `answer` | string | Gold answer |
152
  | `reasoning_level` | int | 1, 2 or 3 |
153
  | `source_chaining` | string | `Homo` or `Hetero` |
154
+ | `sources` | list[string] | Gold evidence sources in hop order (see *Source names*) |
155
  | `evidence_modality` | string | `sql-only`, `text-only` or `composite` |
156
  | `composite_family` | string or null | `entity_linked` or `answer_linked`; null for Level 1 |
157
  | `operation_type` | string | Operation of the final hop (`search`, `aggregate`, `filter`, `rank`, `calculate`, `compare`) |
 
220
 
221
  The [code repository](https://github.com/DISL-Lab/RetailBridge) contains RetailRAG and
222
  the evaluation script. Answer accuracy is judged by an LLM (Muse Ember) against the gold
223
+ answer; results for 16 LLM backbones are on the
224
  [project page](https://retailbridge-kdd.github.io).
225
 
226
  ## License