Dear Michael
Since I have been kind of "at it" on NLP, I thought to be fair to not miss out on this call, I hope similar calls could find these remarks useful:
i. NLP as CHI (Computer-Human Interaction) is fine; ii. some information can be extracted based on text, not all; (there might also be a "culture/disciplinary/technical gap" on the term "meaning";) iii. some or a few text-based phenomena can be captured in statistical/machine learning (if one knows what and how --- without tokenization and with full data (i.e. without discarding any data) --- this can be considered advanced research or the next goalpost compared to what has been customarily done in "NLP"). Most text-based phenomena, claimed in "NLP" thus far, may not really need ML or be compatible with ML, esp. if there are no statistical correlates.
I suppose NLP is still an explicit re-evaluation phase. There is much to clean up. It is important to examine more carefully the interaction between statistical/numerical and textual values.
Thanks for bearing with me and my remarks here.
Best Ada
On Mon, Aug 28, 2023 at 11:08 AM Michael Roth (Saarland University) via Corpora corpora@list.elra.info wrote:
The independent research group on
"Computational Models of Misunderstanding for Complex Instructional Text"
invites applications for one research associate. The position is funded through a grant in the Emmy Noether Programme of the German Research Foundation (DFG---Deutsche Forschungsgemeinschaft), which funds projects similar to an ERC Starting Grant or NSF CAREER Award. The group is headed by Dr. Michael Roth and currently hosted by the Institute for Natural Language Processing ("IMS") at the University of Stuttgart, Germany [1 https://www.ims.uni-stuttgart.de/en/institute/researchgroups/mist/].
The project is concerned with the systematic analysis and computational modelling of text passages that can lead to misunderstandings. A substantial amount of previous work has studied misunderstandings in dialogue, but suitable resources for written language are scarce because misunderstandings cannot be observed directly from a text. Since readers and writers typically do not interact, it is important for authors to ensure that texts leave no room for misinterpretation. Otherwise, for example, medical instructions may be followed incorrectly, and route directions may not guide navigators to their desired destination.
The announced position plays a key role in the project's final phase, leveraging previously created resources (e.g. [2 https://aclanthology.org/2022.lrec-1.354/,3 https://aclanthology.org/2020.lrec-1.702/]) and connecting to the group's award-winning earlier work (e.g. [4 https://aclanthology.org/2021.eacl-srw.5/,5 https://aclanthology.org/2022.semeval-1.146/]). Potential areas of focus for the successful candidate include delving deeper into specific linguistic factors that may lead to misunderstandings (such as elements of implicit or underspecified language), enhancing classification models by incorporating additional information (such as commonsense knowledge or multi-modal context), and/or testing these models in practical applications (such as question answering or machine translation). The position is initially available until February 2025, with a start date as soon as possible (e.g. December 2023) and the possibility of extension (for a total of at least 2 years). Compensation will be in accordance with the German TV-L E13 salary scale at 100% (approx. 4,000 EUR *gross* per month).
Successful applicants will have obtained a Ph.D. (or are close to completing their thesis) in computational linguistics, machine learning, or a closely related field, with a particular interest in semantics and pragmatics or downstream applications. Programming skills and the ability to work in a team are taken for granted. The candidate should be able to work and communicate in English (no proficiency of German is required). Applications should include a motivation letter including research interests, a CV, a list of publications and contact information of up to three references. Applications should be sent *as a single PDF file* to Michael Roth by email. Applications received by 23 September 2023 will receive full consideration, but the position will remain open until filled.
Candidates who identify as female, LGBTQ+ and/or as members of any underrepresented group are particularly encouraged to apply. Feel free to contact Michael Roth (head of group) or Nicola Fanton (PhD student) for any question regarding the group or position.
[1] https://www.ims.uni-stuttgart.de/en/institute/researchgroups/mist/ [2] https://aclanthology.org/2022.lrec-1.354/ [3] https://aclanthology.org/2020.lrec-1.702/ [4] https://aclanthology.org/2021.eacl-srw.5/ (best student paper) [5] https://aclanthology.org/2022.semeval-1.146/ (best task description paper)
-- Dr. Michael Roth Emmy Noether Group Leader Institute for Natural Language Processing University of Stuttgart
Corpora mailing list -- corpora@list.elra.info https://list.elra.info/mailman3/postorius/lists/corpora.list.elra.info/ To unsubscribe send an email to corpora-leave@list.elra.info