diff --git a/README.md b/README.md
index 1ffdb5c14664c1519ed761194062a032f79b48f1..ee7aedb89c7be427f864c9e6d5119c4cfd0a1f3a 100755
--- a/README.md
+++ b/README.md
@@ -16,7 +16,13 @@
 - [License](#license)
 
 ## Updates
+
+2021.9.13:
+
+- Add [MultiWOZ 2.3](https://github.com/lexmen318/MultiWOZ-coref) dataset in `data` dir. The dataset adds co-reference annotations in addition to corrections of dialogue acts and dialogue states. [[paper]](https://arxiv.org/abs/2010.05594)
+
 2021.6.18:
+
 - Add [LAUG](https://github.com/thu-coai/LAUG), an open-source toolkit for Language understanding AUGmentation. It is an automatic method to approximate the natural perturbations to existing data. Augmented data could be used to conduct black-box robustness testing or enhancing training. [[paper]](https://arxiv.org/abs/2012.15262)
 - Add [SC-GPT](https://github.com/pengbaolin/SC-GPT) for NLG. [[paper]](https://arxiv.org/abs/2002.12328)
 
diff --git a/data/multiwoz2.3/MultiWOZ2_3.zip b/data/multiwoz2.3/MultiWOZ2_3.zip
new file mode 100755
index 0000000000000000000000000000000000000000..b1e35eb63e07da9116940ae52adff79f9900da79
Binary files /dev/null and b/data/multiwoz2.3/MultiWOZ2_3.zip differ
diff --git a/data/multiwoz2.3/README.md b/data/multiwoz2.3/README.md
new file mode 100755
index 0000000000000000000000000000000000000000..5cbd4ba29a2c705f41188b0d277368462037dda1
--- /dev/null
+++ b/data/multiwoz2.3/README.md
@@ -0,0 +1,89 @@
+# MultiWOZ 2.3
+
+MultiWOZ 2.3 adds co-reference annotations in addition to corrections of dialogue acts and dialogue states. Please refer to the following paper to get more details:
+**MultiWOZ 2.3: A multi-domain task-oriented dialogue dataset enhanced with annotation corrections and co-reference annotation**. [[PDF]](https://arxiv.org/abs/2010.05594) (Jun. 14, 2021 updated)
+
+Appendices [[PDF]](https://github.com/lexmen318/MultiWOZ-coref/blob/main/appendix_new.pdf) for MultiWOZ 2.3 accepted by NLPCC 2021
+
+If you find our dataset useful and use in your work, please cite the following paper. The bibtex is listed below
+<pre>
+@article{han2020multiwoz,
+  title={MultiWOZ 2.3: A multi-domain task-oriented dialogue dataset enhanced with annotation corrections and co-reference annotation},
+  author={Han, Ting and Liu, Ximing and Takanobu, Ryuichi and Lian, Yixin and Huang, Chongxuan and Wan, Dazhen and Peng, Wei and Huang, Minlie},
+  journal={arXiv preprint arXiv:2010.05594},
+  year={2020}
+}
+</pre>
+
+## Dataset
+
+Three files are included in the zip file:
+
+1. **data.json**: the updated dataset, we add co-reference annotations.
+2. **dialogue_acts.json**: the updated dialogue acts. 
+3. **ontology.json**: the ontology is based on MultiWOZ 2.1 and the only difference is slot format, from domain-semi-slot to domain-slot.
+
+All files have similar format as those of previous datasets (https://github.com/budzianowski/multiwoz).
+
+Except for the corrected and co-reference annotations, we also made the following improvements:
+
+1. The field of "turn_id" is added to all utterances so that they could be referred in co-reference annotations.
+2. There are five dialogues having no "dialogue_act" annotations in MultiWOZ 2.1. These dialogues are annotated manually one by one in MultiWOZ-coref 
+3. Fixed some garbage characters inside MultiWOZ 2.1
+4. The field of "new_goal" is added to all dialogues. The new goal annotations are extracted from the goal descriptions. Note that, "book" and "fail_book" ("info" and "fail_info") are merged, and redundant domains are removed.
+
+```json
+// goal
+{
+    "restaurant": {
+        "fail_book": {
+            "time": "18:30"
+        },
+        "book": {
+            "time": "17:30",
+            "people": "8"
+        }
+    },
+    "train": {}
+}
+```
+
+```json
+// new_goal
+{
+    "restaurant": {
+        "book": {
+            "time": ["18:30", "17:30"],
+            "people": ["8"]
+        }
+    }
+}
+```
+
+
+
+## Experiment
+
+The two models, SUMBT and TRADE, used in the experiment of the paper can be accessed through following links:
+
+SUMBT: https://github.com/SKTBrain/SUMBT <br/>
+TRADE: https://github.com/jasonwu0731/trade-dst <br/>
+
+Please use the scripts provided by the two models to format the data appropriately before you run the models. The ontology comes with MultiWOZ 2.3 is based on the version in MultiWOZ 2.1 and can be directly used for the above two models. The only difference is the format of slot names. Please note that you can freely build up your own ontology.
+
+## More experiments
+On availability, we test our dataset on different DST models. Process scripts for different DST models remain unchanged and are available from their githubs (click the model name if the githubs are still accessible).
+
+| DST Model | MultiWOZ 2.1 | MultiWOZ 2.2 | MultiWOZ 2.3 |
+| --------- | ------------ | ------------ | ------------ |
+| [TRADE](https://github.com/jasonwu0731/trade-dst) | 46.0% | 45.4% | 49.2% |
+| [SUMBT](https://github.com/SKTBrain/SUMBT) | 49.2% | 49.7% | 52.9% |
+| [COMER](https://github.com/renll/ComerNet) | 48.8% | -- | 50.2% |
+| [DSTQA](https://github.com/alexa/dstqa) | 51.2% | --| 51.8% |
+| [SOM-DST](https://github.com/clovaai/som-dst) | 53.1% | -- | 55.5% |
+| [TripPy](https://gitlab.cs.uni-duesseldorf.de/general/dsml/trippy-public) | 55.3% |  -- | 63.0% |
+| [SimpleTOD*](https://github.com/salesforce/simpletod) | 50.3% (55.7%) | -- | 51.3% |
+| [ConvBERT-DG-Multi](https://github.com/alexa/dialoglue) | 58.7% | -- | 67.9% |
+| [SAVN](https://github.com/wyxlzsq/savn) | 54.5% | -- | 58.0% |
+
+Please note that "--" means that no performence reported. * in SimpleTOD means that we only run the code for DST by keeping `dontcare` and `none`. For further details, please refer to the github: https://github.com/salesforce/simpletod.
diff --git a/data/multiwoz2.3/appendix_new.pdf b/data/multiwoz2.3/appendix_new.pdf
new file mode 100755
index 0000000000000000000000000000000000000000..0ba95814e4a783301e95099935134695ced2331e
Binary files /dev/null and b/data/multiwoz2.3/appendix_new.pdf differ