diff --git a/content/authors/alumnus/_index.md b/content/authors/alumnus/_index.md
index 6d11b00..20642e7 100644
--- a/content/authors/alumnus/_index.md
+++ b/content/authors/alumnus/_index.md
@@ -26,6 +26,5 @@ highlight_name: false
 user_groups:
   - Alumni
 
----
-
+---xw
 An alumnus
\ No newline at end of file
diff --git a/publications.bib b/publications.bib
new file mode 100644
index 0000000..3a8ab8a
--- /dev/null
+++ b/publications.bib
@@ -0,0 +1,27 @@
+@inproceedings{huang-etal-2022-lightweight,
+    title = "Lightweight Contextual Logical Structure Recovery",
+    author = "Huang, Po-Wei  and
+      Ramesh Kashyap, Abhinav  and
+      Qin, Yanxia  and
+      Yang, Yajing  and
+      Kan, Min-Yen",
+    editor = "Cohan, Arman  and
+      Feigenblat, Guy  and
+      Freitag, Dayne  and
+      Ghosal, Tirthankar  and
+      Herrmannova, Drahomira  and
+      Knoth, Petr  and
+      Lo, Kyle  and
+      Mayr, Philipp  and
+      Shmueli-Scheuer, Michal  and
+      de Waard, Anita  and
+      Wang, Lucy Lu",
+    booktitle = "Proceedings of the Third Workshop on Scholarly Document Processing",
+    month = oct,
+    year = "2022",
+    address = "Gyeongju, Republic of Korea",
+    publisher = "Association for Computational Linguistics",
+    url = "https://aclanthology.org/2022.sdp-1.5",
+    pages = "37--48",
+    abstract = "Logical structure recovery in scientific articles associates text with a semantic section of the article. Although previous work has disregarded the surrounding context of a line, we model this important information by employing line-level attention on top of a transformer-based scientific document processing pipeline. With the addition of loss function engineering and data augmentation techniques with semi-supervised learning, our method improves classification performance by 10{\%} compared to a recent state-of-the-art model. Our parsimonious, text-only method achieves a performance comparable to that of other works that use rich document features such as font and spatial position, using less data without sacrificing performance, resulting in a lightweight training pipeline.",
+}