Our goal is to build a computational model of the process of sentence comprehension in a text passage, using text corpora and other linguistic resources, and validate it using behavioral and brain imaging data.
The main role of the intern will be to help in developing and implementing methods to extract information from text corpora, as well as models for emulating human performance on a variety of benchmark psychological tests. Examples of tasks you might be asked to do:
- generate distributed semantic representations for words/concepts, sentences and passages - extract and combine structured information from text corpora or resources such as WordNet, FrameNet and various knowledge bases - prepare and process text corpora (parse, convert to other formats, etc) - implement evaluation tasks to benchmark the models developed
It is an unusual position in that you would be doing machine learning tasks to further cognitive neuroscience goals, gaining experience in both areas. In addition to the core research, we will also be delivering a system to the funding agency; hence, this is a fast-paced project with multiple opportunities for publication. This work is being carried out in collaboration with researchers at MIT, MGH and Princeton University and funded by the IARPA Knowledge Representation in Neural Systems program ( http://www.iarpa.gov/index.php/research-programs/krns ).
Requirements: - current student in a graduate program in Computer Science/Electrical Engineering/Machine Learning - experience developing software in MATLAB, Python or Perl - experience in Computational Linguistics and NLP tasks (e.g. using frameworks such as CoreNLP or NLTK) - availability for at least 3 months (a longer period is possible, and the timeline is flexible)
If interested, please send your resume to francisco-pereira at siemens.com (your email can serve as a cover letter, if there is anything you would like to elaborate on).
thank you! Francisco -------------- next part -------------- A non-text attachment was scrubbed... Name: not available Type: text/html Size: 2630 bytes Desc: not available URL: <https://mailman.uib.no/public/corpora/attachments/20160310/912cda7b/attachment.txt>