<span style="font-family:arial, sans-serif;font-size:13px;border-collapse:collapse">Dear all,<br><br>I'm happy to announce the release of What's Wrong With My NLP? 0.2.3:<br>
a visualizer and graphical diff for NLP problems, such as </span><div><span style="font-family:arial, sans-serif;font-size:13px;border-collapse:collapse"><br></span></div><div><span style="font-family:arial, sans-serif;font-size:13px;border-collapse:collapse">* dependency parsing, </span></div>
<div><span style="font-family:arial, sans-serif;font-size:13px;border-collapse:collapse">* semantic role labeling, </span></div><div><span style="font-family:arial, sans-serif;font-size:13px;border-collapse:collapse">* chunking</span></div>
<div><span style="font-family:arial, sans-serif;font-size:13px;border-collapse:collapse">* NER</span></div><div><span style="font-family:arial, sans-serif;font-size:13px;border-collapse:collapse">* event extraction</span></div>
<div><span style="font-family:arial, sans-serif;font-size:13px;border-collapse:collapse">* multilingual alignments</span></div><div><span style="font-family:arial, sans-serif;font-size:13px;border-collapse:collapse">* constituent trees</span></div>
<div><span style="font-family:arial, sans-serif;font-size:13px;border-collapse:collapse">* anything you can visualize with spans over tokens, and edges between tokens.</span></div><div><font class="Apple-style-span" face="arial, sans-serif"><span class="Apple-style-span" style="border-collapse: collapse;"><br>
</span></font></div><div><font class="Apple-style-span" face="arial, sans-serif"><span class="Apple-style-span" style="border-collapse: collapse;">It supports diff'ing of gold and guess solutions, export to EPS, keyword search through corpora etc. </span></font></div>
<div><font class="Apple-style-span" face="arial, sans-serif"><span class="Apple-style-span" style="border-collapse: collapse;"><br></span></font></div><div><font class="Apple-style-span" face="arial, sans-serif"><span class="Apple-style-span" style="border-collapse: collapse;">What's new in this version: it allows you provide additional information about edges and spans in your predictions. For example, for first order dependency parsing models you can show feature vectors associated with false positive or negative edges. I've found this to be a very helpful feature when debugging probabilistic models of language. For info on how to use this go to: <a href="http://code.google.com/p/whatswrong/wiki/GenericFormat">http://code.google.com/p/whatswrong/wiki/GenericFormat</a></span></font></div>
<div><font class="Apple-style-span" face="arial, sans-serif"><span class="Apple-style-span" style="border-collapse: collapse;"><br></span></font></div><div><span class="Apple-style-span" style="font-family: arial, sans-serif; font-size: 13px; border-collapse: collapse; ">The software is written in Java and licensed under the GPL v3. To</span></div>
<div><span style="font-family:arial, sans-serif;font-size:13px;border-collapse:collapse">download, go to<br><br><a href="http://whatswrong.googlecode.com" style="color:rgb(0, 84, 136)" target="_blank">http://<span style="background-color:rgb(255, 255, 204);background-repeat:initial initial">whatswrong</span>.googlecode.com</a><br>
<br>I'd be happy about any kind of feedback,</span>
</div><div><span style="font-family:arial, sans-serif;font-size:13px;border-collapse:collapse">Sebastian</span></div>