link add

Author: postylem

_bibliography/pubs.bib

@@ -4,8 +4,8 @@ @misc{hoover-etal-2022-amlap
   author       = {Hoover, Jacob Louis and Sonderegger, Morgan and Piantadosi, Steven T. and O'Donnell, Timothy J.},
   year         = {2022},
   howpublished = {Poster at Architectures and  Mechanisms for Language Processing (AMLaP 28)},
-  day          = 6,
-  month        = sep,
+  day          = {6},
+  month        = {September},
   address      = {{York, England}},
   url          = {https://virtual.oxfordabstracts.com/#/event/3067/submission/297},
   poster       = {https://d3ijlhudpq9yjw.cloudfront.net/cdd9fcfe-abaa-4d50-b0a0-9723ab9e9bc9.pdf}
@@ -15,7 +15,7 @@ @article{hoover-etal-2023-plausibility
   title      = {The Plausibility of Sampling as an Algorithmic Theory of Sentence Processing},
   author     = {Hoover, Jacob Louis and Sonderegger, Morgan and Piantadosi, Steven T. and O'Donnell, Timothy J.},
   year       = {2023},
-  month      = jul,
+  month      = {July},
   journal    = {Open Mind: Discoveries in Cognitive Science},
   volume     = {7},
   pages      = {350--391},
@@ -57,7 +57,7 @@ @inproceedings{hoover-etal-2021-emnlp
   author    = {Hoover, Jacob Louis and Du, Wenyu and Sordoni, Alessandro and O{'}Donnell, Timothy J.},
   booktitle = {Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing},
   code      = {https://github.com/mcqll/cpmi-dependencies},
-  month     = nov,
+  month     = {November},
   pages     = {2941--2963},
   poster    = {2021.10.11.EMNLP.poster.pdf},
   publisher = {Association for Computational Linguistics},
@@ -72,9 +72,10 @@ @thesis{hoover.j:2024phd
   author = {Hoover, Jacob Louis},
   type   = {PhD},
   year   = {2024},
-  month  = aug,
+  month  = {August},
   note   = {PhD Thesis, McGill University, Linguistics Department},
   school = {McGill University},
-  langid = {english},
-  url    = {https://jahoo.github.io/assets/dissertation.pdf}
+  langid = {en-CA},
+  url    = {https://escholarship.mcgill.ca/concern/theses/r494vr42w},
+  pdf    = {https://jahoo.github.io/assets/dissertation.pdf}
 }

_config.yml

@@ -27,7 +27,7 @@ scholar:
   order: descending
   # group_by: year
   # group_order: descending
-  bibtex_skip_fields: [ "abstract", "month_numeric", "poster", "handout", "pdf", "code", "slides" ]
+  bibtex_skip_fields: ["abstract", "code", "handout", "month_numeric", "myurl", "myurltitle", "openaccess", "pdf", "poster", "preprint", "slides"]
   # type_names: { article: Papers }
   # linebreaks: true # 
   source: /_bibliography

_posts/2020-12-22-training-tensor-trains.md

@@ -1,6 +1,6 @@
 ---
 layout: post
-title: Training Tensor Trains
+title: A practical comparison of tensor train models
 comments: true
 published: true 
 tags:
@@ -13,12 +13,12 @@ I worked on a project with Jonathan Palucci exploring the trainability of a cert
 <div style="text-align: center;"><img width="400" src="/assets/2020-12-22-training-tensor-trains-fig2.png"></div>
 
 
-There is a general correspondence between tensor networks and graphical models, and in particular, when restricted to non-negative valued parameters, Matrix Product States are equivalent to Hidden Markov Models. [Glasser _et al_. 2019](https://arxiv.org/abs/1907.03741) discussed this correspondence, and proved theoretical results about these non-negative models, as well as similar real-- and complex--valued tensor trains.  They supplemented their theoretical results with evidence from numerical experiments.  In this project, we re-implemented models from their paper, and also implemented time-homogeneous versions of their models.
+There is a general correspondence between tensor networks and graphical models, and in particular, when restricted to non-negative valued parameters, [Matrix Product States](https://tensornetwork.org/mps/) are equivalent to Hidden Markov Models (HMMs)). [Glasser _et al_. 2019](https://arxiv.org/abs/1907.03741) discussed this correspondence, and proved theoretical results about these non-negative models, as well as similar real-- and complex--valued tensor trains.  They supplemented their theoretical results with evidence from numerical experiments.  In this project, we re-implemented models from their paper, and also implemented time-homogeneous versions of their models.
 We replicated some of their results for non-homogeneous models, adding a comparison with homogeneous models on the same data.  We found evidence that homogeneity decreases ability of the models to fit non-sequential data, but preliminarily observed that on sequential data (for which the assumption of homogeneity is justified), homogeneous models achieved an equally good fit with far fewer parameters. Surprisingly, we also found that the more powerful non time-homogeneous positive MPS performs identically to a time homogeneous HMM.
 
 📊 Poster --> [here (PDF)](/assets/pdfs/2020.12.15.tensor-trains-poster.pdf).  
 
-📄 Writeup --> [here (PDF)](/assets/pdfs/2020.12.22.tensor-trains-writeup.pdf).  
+📄 Writeup titled _A practical comparison of tensor train models: The effect of homogeneity_ --> [here (PDF)](/assets/pdfs/2020.12.22.tensor-trains-writeup.pdf).  
 
-💻 Code --> [on GitHub](https://github.com/postylem/tensor_network_project).
+💻 Code --> [on GitHub](https://github.com/postylem/comparison-of-tensor-train-models).