<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1">
<style type="text/css" style="display:none;"> P {margin-top:0;margin-bottom:0;} </style>
</head>
<body dir="ltr">
<div style="font-family: Calibri, Arial, Helvetica, sans-serif; font-size: 12pt; color: rgb(0, 0, 0);" class="elementToProof">
<br>
</div>
<span class="x_elementToProof" style="font-size:12pt;margin:0px;background-color:rgb(255, 255, 255)"><span class="x_x_elementToProof x_ContentPasted0 ContentPasted0" style="margin:0px;background-color:rgb(255, 255, 255)">Dear all,</span></span>
<div class="x_elementToProof" style="font-size:12pt;margin:0px;background-color:rgb(255, 255, 255)">
<div class="x_x_elementToProof" style="margin:0px;background-color:rgb(255, 255, 255)">
<br class="x_ContentPasted0 ContentPasted0">
</div>
<div class="x_x_elementToProof x_ContentPasted0 ContentPasted0" style="margin:0px;background-color:rgb(255, 255, 255)">
For my thesis, I will try to offer Haskell coding assignments for a course in a more dynamic way, where the next assignment is based on the knowledge gap of the learner. For this, I will use a neural network, which will have to process a submission of a learner.
 Therefore, I want to use a language model that extracts features from the submission. </div>
<div class="x_x_elementToProof x_ContentPasted0 ContentPasted0" style="margin:0px;background-color:rgb(255, 255, 255)">
I was planning to use BERT in a first phase, since this model creates embeddings from a document. However, this model is not trained on any programming language, so this may turn out to be insufficient. Moreover, models such as CodeBERT are trained on source
 code, but only on very well known programming languages.</div>
<div class="x_x_elementToProof" style="margin:0px;background-color:rgb(255, 255, 255)">
<br class="x_ContentPasted0 ContentPasted0">
</div>
<div class="x_x_elementToProof x_ContentPasted0 ContentPasted0" style="margin:0px;background-color:rgb(255, 255, 255)">
I was wondering if someone had an idea if there already exists a language model that can create meaningful representations for Haskell. More general models such as Codex (paying API) or GPT-Code-Clippy (free and opensource) could also be tried out, but if there
 exists a model that has been trained on Haskell specific data, that would of course be the most beneficial option.</div>
<div class="x_x_elementToProof" style="margin:0px;background-color:rgb(255, 255, 255)">
<span style="margin:0px"><br class="x_ContentPasted0 ContentPasted0">
</span></div>
<div class="x_x_elementToProof" style="margin:0px;background-color:rgb(255, 255, 255)">
<span class="x_ContentPasted0 ContentPasted0" style="margin:0px">If you want to read more about my thesis, you can read more about it in the attachment.</span><br class="x_ContentPasted0 ContentPasted0">
</div>
<div class="x_x_elementToProof" style="margin:0px;background-color:rgb(255, 255, 255)">
<span style="margin:0px"><br class="x_ContentPasted0 ContentPasted0">
</span></div>
<div class="x_x_elementToProof" style="margin:0px;background-color:rgb(255, 255, 255)">
<span class="x_ContentPasted0 ContentPasted0" style="margin:0px">Thanks in advance!</span></div>
<div class="x_x_elementToProof" style="margin:0px;background-color:rgb(255, 255, 255)">
<br class="x_ContentPasted0 ContentPasted0">
</div>
<div class="x_x_elementToProof x_ContentPasted0 ContentPasted0" style="margin:0px;background-color:rgb(255, 255, 255)">
Kind regards,</div>
<div class="x_x_elementToProof x_ContentPasted0 ContentPasted0" style="margin:0px;background-color:rgb(255, 255, 255)">
Jorrit Willaert<br class="x_ContentPasted0 ContentPasted0">
</div>
<div class="x_x_elementToProof x_x_ContentPasted0" style="margin:0px;background-color:rgb(255, 255, 255)">
<a href="https://www.linkedin.com/in/jorrit-willaert-166a64198/" target="_blank" rel="noopener noreferrer" data-auth="NotApplicable" data-safelink="true" data-linkindex="0" class="x_ContentPasted0 ContentPasted0" style="margin:0px">https://www.linkedin.com/in/jorrit-willaert-166a64198/</a><br class="x_ContentPasted0 ContentPasted0">
</div>
</div>
<span class="x_elementToProof" style="font-size:12pt;margin:0px;background-color:rgb(255, 255, 255)"><span class="x_x_elementToProof x_x_ContentPasted0 x_x_ContentPasted1" style="margin:0px;background-color:rgb(255, 255, 255)"><a href="https://github.com/JorritWillaert" target="_blank" rel="noopener noreferrer" data-auth="NotApplicable" data-safelink="true" data-linkindex="1" class="x_ContentPasted0 ContentPasted0" style="margin:0px">https://github.com/JorritWillaert</a></span></span>
<div class="elementToProof">
<div id="Signature">
<div>
<div name="divtagdefaultwrapper" style="font-family:Calibri,Arial,Helvetica,sans-serif; font-size:; margin:0">
<div></div>
</div>
</div>
</div>
</div>
</body>
</html>