From 95dbee642e33b08d29166b279d493c96cbd9d1cd Mon Sep 17 00:00:00 2001
From: Guillaume Dalle <22795598+gdalle@users.noreply.github.com>
Date: Thu, 22 Feb 2024 16:48:29 +0100
Subject: [PATCH] Fix biblio

---
 paper/HMM.bib  | 27 ++++++++++++++-------------
 paper/paper.md |  4 ++--
 2 files changed, 16 insertions(+), 15 deletions(-)

diff --git a/paper/HMM.bib b/paper/HMM.bib
index 87d81d48..e3bb7639 100644
--- a/paper/HMM.bib
+++ b/paper/HMM.bib
@@ -80,6 +80,18 @@ @book{cappeInferenceHiddenMarkov2005
   file = {/home/gdalle/snap/zotero-snap/common/Zotero/storage/2HYZE7ZD/Cappé et al_2005_Inference in Hidden Markov Models.pdf;/home/gdalle/snap/zotero-snap/common/Zotero/storage/QRNV9CL8/Cappé et al. - 2006 - Inference in Hidden Markov Models.pdf}
 }
 
+@software{changDynamaxStateSpace2024,
+  title = {Dynamax:  {{State Space Models}} Library in {{JAX}}},
+  author = {Chang, Peter and Harper-Donnelly, Giles and Kara, Aleyna and Li, Xinglong and Linderman, Scott and Murphy, Kevin},
+  date = {2024-02-22T04:10:59Z},
+  origdate = {2022-04-11T23:42:29Z},
+  url = {https://github.com/probml/dynamax},
+  urldate = {2024-02-22},
+  abstract = {State Space Models library in JAX},
+  organization = {{Probabilistic machine learning}},
+  keywords = {hmm}
+}
+
 @thesis{dalleMachineLearningCombinatorial2022,
   type = {phdthesis},
   title = {Machine Learning and Combinatorial Optimization Algorithms, with Applications to Railway Planning},
@@ -97,9 +109,9 @@ @thesis{dalleMachineLearningCombinatorial2022
   file = {/home/gdalle/snap/zotero-snap/common/Zotero/storage/CEVJMUP4/Dalle - Machine learning and combinatorial optimization al.pdf}
 }
 
-@software{hmmlearnHmmlearnHiddenMarkov2023,
+@software{hmmlearndevelopersHmmlearnHiddenMarkov2023,
   title = {Hmmlearn: {{Hidden Markov Models}} in {{Python}}, with Scikit-Learn like {{API}}},
-  author = {{hmmlearn}},
+  author = {{hmmlearn developers}},
   date = {2023},
   url = {https://github.com/hmmlearn/hmmlearn},
   urldate = {2023-09-12},
@@ -139,17 +151,6 @@ @unpublished{ondelGPUAcceleratedForwardBackwardAlgorithm2021
   file = {/home/gdalle/snap/zotero-snap/common/Zotero/storage/XRKC5QBG/Ondel et al. - 2021 - GPU-Accelerated Forward-Backward Algorithm with Ap.pdf}
 }
 
-@software{ProbmlDynamax2024,
-  title = {Probml/Dynamax},
-  date = {2024-02-22T04:10:59Z},
-  origdate = {2022-04-11T23:42:29Z},
-  url = {https://github.com/probml/dynamax},
-  urldate = {2024-02-22},
-  abstract = {State Space Models library in JAX},
-  organization = {{Probabilistic machine learning}},
-  keywords = {hmm}
-}
-
 @article{qinDirectOptimizationApproach2000,
   title = {A {{Direct Optimization Approach}} to {{Hidden Markov Modeling}} for {{Single Channel Kinetics}}},
   author = {Qin, Feng and Auerbach, Anthony and Sachs, Frederick},
diff --git a/paper/paper.md b/paper/paper.md
index 8e94df1d..5d6c9364 100644
--- a/paper/paper.md
+++ b/paper/paper.md
@@ -48,8 +48,8 @@ In this industrial use case, the observations were marked temporal point process
 
 Unfortunately, nearly all implementations of HMMs we surveyed (in Julia and Python) expect the observations to be generated by a _predefined set of distributions_, with _no control dependency_.
 In Julia, the reference package `HMMBase.jl` [@mouchetHMMBaseJlHidden2023] requires compliance with the `Distributions.jl` [@besanconDistributionsJlDefinition2021] interface, which precludes anything not scalar- or array-valued, let alone point processes.
-In Python, `hmmlearn` [@hmmlearnHmmlearnHiddenMarkov2023], `pomegranate` [@schreiberPomegranateFastFlexible2018a] each offer a catalogue of discrete and continuous distributions, but do not allow for easy extension by the user.
-The more recent `dynamax` [@ProbmlDynamax2024] is the only package adopting an extensible interface with optional controls, similar to ours.
+In Python, `hmmlearn` [@hmmlearndevelopersHmmlearnHiddenMarkov2023], `pomegranate` [@schreiberPomegranateFastFlexible2018] each offer a catalogue of discrete and continuous distributions, but do not allow for easy extension by the user.
+The more recent `dynamax` [@changDynamaxStateSpace2024] is the only package adopting an extensible interface with optional controls, similar to ours.
 
 Focusing on Julia specifically, other downsides of `HMMBase.jl` include the lack of support for _multiple observation sequences_ or _sparse transition matrices_, and the mandatory use of _64-bit floating point numbers_.
 Two other packages provide functionalities that `HMMBase.jl` lacks: `HMMGradients.jl` [@antonelloHMMGradientsJlEnables2021] contains a _differentiable loglikelihood function_, while `MarkovModels.jl` [@ondelGPUAcceleratedForwardBackwardAlgorithm2021] focuses on GPU acceleration.