the programme committee has completed their hard work of reviewing submissions; each team will receive at least five reviews (of which two by track chairs, though reviewer identities remain anonymous).
the track chairs plan on announcing their decisions before noon tomorrow (monday, may 27). each team is then asked to prepare their camera-ready, non-anonymous manuscript and upload the final version to EasyChair no later than the following monday, june 2.
on tuesday, june 11, we will hold the actual workshop: 14:15-16:00 in seminar room perl (OJD, 2453). each team will be allocated a total of ten minutes, of which two minutes should be reserved for questions.
with only eight minutes available to present, you will need to focus on high-level questions and a selection of results. seeing as WNNLP 2019 is a gathering of spe...
as we hope all teams are making good progress on their submissions, we would like to hear with everyone about the best possible date for the final, short presentations. so far, we have scheduled WNNLP 2019 for thursday, june 13, i.e. the usual lecture slot.
it turns out, at least one of you will have a hard time presenting their work that day. hence, we are considering holding the workshop already on tuesday, june 11, at 14:15, i.e. the usual laboratory slot. the slightly earlier day would be preferable for at least one student and one instructor.
however, at this point we feel we can only reschedule the workshop if the move would not inconvenience anyone. thus, we tentatively propose holding WNNLP 2019 on tuesday, june 11, but would kindly ask that anyone who would have a conflict that day to email us (...
The exam is on now: we have published sets of instructions, two to three pages each, for the four tracks; 14 students (in ten teams) have declared their task (with a neat distribution of four, three, two, and one teach each for NER, NLI, NOT, and RC, respectively). The instructions and supporting data and code, where applicable, are available to course participants in the IN5550 Microsoft GitHub service at UiO.
Please consult the main web page for the WNNLP 2019 workshop (aka the exam) for additional instructions, and please monitor that page regularly as we will be posting updates there. Recall that the primary goal of the home exam is to prepare a scientific paper, grounded in your experiments, so please do not leave the writing until the very end. The deadline for paper submissions will b...
The third obligatory assignment (Sentence-level Sentiment Classification with Convolutional Neural Networks) is now out. The deadline for submissions is April 5th.
We have just published additional information on the final exam for the class, which will take the form of a 'baby' research period, technically a home exam in the period May 2 to May 16, 2019. To allow a little extra time for completion of the third and final obligatory assignment before the Easter break, we have extended the submission deadline for Assignment (3) to April 5, 2018.
The second obligatory assignment is now published.
You can start working on it right away. The deadline for submissions is March 8.
The obligatory assignment 2 publication is postponed a bit. We hope it to go public on February 21 or February 22 at the latest.
Unfortunately, there will be no group session on February 5, due to all teachers leaving for the NLPL winter school.
Because of that, the Obligatory 1 submission deadline is postponed to February 15.
On February 12, there will be a regular group session specifically devoted to obligatory 1.
Correspondingly, all the subsequent assignments are shifted a week forward (see the table).
The first obligatory assignment is now published. You are free to start working on it. However (especially if you lack the machine learning background) it is advised to wait until the next lecture on January 31.
Today (Tuesday, January 22), Vinit (Chief Nerd), Andrey (Associate Nerd), and Stephan (Assistant Nerd) will hold our first laboratory session. The game plan for today is to demonstrate how to develop (in Python) on Abel, run batch jobs through the queuing system, edit and debug your code effectively, ‘vectorize’ a document, and refresh our understanding of NumPy and basic linear algebra on vectors and matrics. Welcome!
IN5550 is a new course at IFI with a near-exclusive focus on applications of deep neural networks to potentially large volumes of natural language data. The first lecture tomorrow (Thursday, January 17) will provide a high-level introduction to the topic and review all relevant course ‘mechanics’. To gauge participant background, we ask everyone to submit a brief and anonymous on-line questionnaire (we will briefly review results towards the end of the lecture tomorrow).