diff --git a/algorithms/tang.yaml b/algorithms/tang.yaml
new file mode 100644
index 0000000..8b55903
--- /dev/null
+++ b/algorithms/tang.yaml
@@ -0,0 +1,70 @@
+---
+# Description of a seizure detection algorithm
+title: "Automated Seizure Detection using Transformer Models on Multi-Channel EEGs" 
+image: "ghcr.io/esl-epfl/tang_2021:latest"
+authors:
+  - given-names: Siyi
+    family-names: Tang
+    affiliation: Stanford Universit
+  - given-names: Clément
+    family-names: Samanos
+    affiliation: EPFL-ESL
+  - given-names: Jonathan
+    family-names: Dan
+    email: jonathan.dan@epfl.ch
+    affiliation: EPFL-ESL
+    orcid: 'https://orcid.org/0000-0002-2338-572X'
+version: 0.1
+date-released: "2022-04-22"
+abstract: >
+  Automated seizure detection and classification from electroencephalography
+  (EEG) can greatly improve seizure diagnosis and treatment. However, several
+  modeling challenges remain unaddressed in prior automated seizure detection
+  and classification studies: (1) representing non-Euclidean data structure in EEGs,
+  (2) accurately classifying rare seizure types, and (3) lacking a quantitative inter-
+  pretability approach to measure model ability to localize seizures. In this study,
+  we address these challenges by (1) representing the spatiotemporal dependencies
+  in EEGs using a graph neural network (GNN) and proposing two EEG graph
+  structures that capture the electrode geometry or dynamic brain connectivity, (2)
+  proposing a self-supervised pre-training method that predicts preprocessed sig-
+  nals for the next time period to further improve model performance, particu-
+  larly on rare seizure types, and (3) proposing a quantitative model interpretabil-
+  ity approach to assess a model's ability to localize seizures within EEGs. When
+  evaluating our approach on seizure detection and classification on a large pub-
+  lic dataset (5,499 EEGs), we find that our GNN with self-supervised pre-training
+  achieves 0.875 Area Under the Receiver Operating Characteristic Curve on seizure
+  detection and 0.749 weighted F1-score on seizure classification, outperforming
+  previous methods for both seizure detection and classification. Moreover, our
+  self-supervised pre-training strategy significantly improves classification of rare
+  seizure types (e.g. 47 points increase in combined tonic seizure accuracy over
+  baselines). Furthermore, quantitative interpretability analysis shows that our GNN
+  with self-supervised pre-training precisely localizes 25.4% focal seizures, a 21.9
+  point improvement over existing CNNs. Finally, by superimposing the identified
+  seizure locations on both raw EEG signals and EEG graphs, our approach could
+  provide clinicians with an intuitive visualization of localized seizure regions.
+  This model was trained on TUH EEG Seizure Corpus v2.0.3
+license: MIT
+repository: https://github.com/esl-epfl/tang_2021
+
+# List all datasets that were used to train this algorithm
+Dataset:
+    - title: "TUH EEG Seizure Corpus v2.0.3"
+      license: "https://isip.piconepress.com/projects/nedc/forms/tuh_eeg.pdf"
+      identifiers:
+        - description: >
+            This database is a subset of the TUH EEG Corpus that was collected 
+            from archival records of clinical EEG at Temple University Hospital 
+            recorded between 2002 – 2017. From this large dataset, a subset of 
+            files with a high likelihood of containing seizures was retained 
+            based on clinical notes and on the output of seizure detection 
+            algorithms.
+            V2.0.0 contains 7377 .edf files from 675 subjects for a total 
+            duration of 1476 hours of data. The files are mostly short 
+            (avg. 10 minutes). The dataset has heterogeneous sampling frequency 
+            and number of channels. All files are acquired at a minimum of 250 
+            Hz. A minimum of 17 EEG channels is available in all recordings. 
+            They are positioned according to the 10-20 system.
+            The annotations are provided as .csv and contain the start time, 
+            stop, channel and seizure type.
+          type: url
+          value: "https://isip.piconepress.com/projects/nedc/html/tuh_eeg/"