Header

Search
Department of Computational Linguistics Digital Linguistics

Student Projects

In our lab we regularly have students that are working on projects or theses with us. There are multiple ways of how you can join our lab as a student. Please see the list below to check out how you can join depending on the project type!

We get many emails and requests. It is not possible to consider everything, especially if emails are messy and important information is missing.

 

Table of contents

Research / Student Assistants

We regularly have open positions for research / student assistants. Assistants help us, for example, to collect data in our lab, complete small tasks for our various projects (e.g., implement web forms for participants to fill in) and other tasks that are related to our daily research life.

If there is no position announced (e.g., via the CL list) you can always express your interest in joining our lab as a research / student assistant by contacting Prof. Dr. Lena A. Jäger.

Make sure to include the following information in your email:

  • What are you interested in?
  • What is your background (e.g., programming skills, experience with data collection, etc.)?
  • Possible start date and duration
  • If you are interested in a tutor / teaching assistance position: please specify whether you want to receive credits or money as a compensation.
Top

Programming Projects

Some university curricula include modules where students have to work on individual projects. For a list of possible thesis projects, see project opportunities. Please note that the list of projects on that page is not exhaustive; we encourage you to approach us with your own research idea.

For more informations regarding programming projects, refer to the department-wide guidelines as well as to our group-specific guidelines below.

Please get in touch with Prof. Dr. Lena A. Jäger in case you are interested in doing your thesis with us (or directly with the respective supervisor(s) if referring to a project listed among the project opportunities).

Please include the following information in your email:

  • What type of project are you looking for?
  • How many credits would you like to earn? This information is important as it defines how many hours you are expected to work on the project.
  • What is the timeframe (start and end date)?
Top

BA and MA thesis projects

We are happy to see that you are interested in writing your thesis with us! For a list of possible thesis projects, see project opportunities. Please note that the list of projects on that page is not exhaustive; we encourage you to approach us with your own research idea. You can find out more about our current interests in the publication or the research section.

For more informations regarding thesis projects, refer to our group-specific guidelines below.

Please get in touch with Prof. Dr. Lena A. Jäger in case you are interested in doing your thesis with us (or directly with the respective supervisor(s) if referring to a project listed among the project opportunities).

Top

Guidelines for theses and projects with the Digital Linguistics Lab

The following guidelines are also available as a PDF (PDF, 165 KB).

Researchers in the Dili Lab supervise different types of study projects that are required or optional when pursuing a degree at the Department of Computational Linguistics Institute (CL, Faculty of Arts and Social Sciences) or the Department of Informatics (IfI, Faculty of Business, Economics and Informatics) at UZH. Please always consult the study regulations provided by the respective department or faculty.

You are responsible for following the departmental guidelines / submission process. The present  guidelines do not replace any guidelines provided by your department or faculty.

General Weekly Meetings

  • When: Weekly, Thursdays 15:00-15:55
  • Where: AND-4.67
  • Why attend:
    • Propose a topic and get it approved (step 2 below).
    • Seeking feedback on your work (step 3 below).
    • Ask any organizational questions.
    • See what other students are working on.
    • For finding a topic, do not inquire via e-mail or Teams, but follow the steps below.

Steps in taking a Programming Project, Master’s Project or any Thesis

  1. Choose a topic.
    • Propose your own project idea, or
    • Select one of our project ideas from the list.
  2. Present your project idea.
    • Present your project proposal in the meeting (see advice below). 
    • For Master’s/ Bachelor’s theses: Agree on a main supervisor. Together with your main supervisor, agree on a title. Download and sign the form "Application for assigning the topic for the final thesis". You must register your thesis before starting to work on your project. 
  3. Get your project approved.
    • Based on the topic you chose, write a proposal and get it approved by your supervisor(s). 
  4. Report on your progress to collect feedback (multiple times).
    • Give multiple progress presentations (see advice below);
    • Before submitting your final report/thesis, send a draft version of your report/thesis (see advice below) to your advisor.
    • Incorporate feedback.
  5. Submission
    • For Bachelor’s theses at CL or Ifi and Master’s theses at CL, please submit your thesis via the interface or process provided by your department.
    • For Ifi Master’s Projects, you will need to give a final presentation AND submit a report/thesis
    • For Ifi Master’s Projects and Ifi MSc theses, you will need to give a final presentation and submit a report
      • Request an appointment for your presentation via your main supervisor.
      • Submit your report at least one week prior to the presentation date.
      • Report details: 
      • Presentation details: 
    • For Ifi MSc theses, you will need to give a final presentation AND submit a thesis
      • Request an appointment for your presentation via your main supervisor.
      • Submit your report at least one week (7 days) prior to the presentation date.
      • Report details: 
      • Presentation details: 
    • For Programming Projects submit your code base (e.g., a GitHub repository) by the submission deadline agreed on with your supervisor. The last commit before the deadline will be considered.

 
General advice on the presentation of your work

Oral Presentations in the Supervision Meeting:  Depending on your topic, please select only the relevant sub-topics on which you would like to get feedback or which are necessary as background information for the audience. Time limit: 15 minutes!

Bachelor’s/Master’s Theses, Project Reports: Please use the structure below as a guideline to present your work.

Proposals for Bachelor’s/Master’s Theses or Projects: Please use the structure below as a guideline to present your work. In case you do not have results yet, you may discuss what results you expect and how you would interpret them.

  • For machine learning projects:
    • Problem setting or research questions:
      • What (real-world) problem is solved using machine learning?
      • Formalize the problem setting, what are inputs and outputs?
      • How can you quantify model performance: Which metrics are suitable?
      • Present your research questions
    • Current State-of-the-Art:
    • Research literature on the stated problem (use Google, Google Scholar, ResearchGate, PubMed etc.)
    • Explain each state-of-the-art-approach, including but not limited to their input and (machine learning) algorithm.
    • Explain the relevant context of your research questions.
    • Your approach/idea/methods:
      • Motivation: e.g. identify shortcomings of existing methods and potential improvements.
      • Overall architecture of your model(s).
      • Methods used and experiments conducted to answer research questions
    • Evaluation protocol:
      • What protocol is or will be used: Cross-validation (recommended!) or Hold-Out, … ?
      • Hyperparameter tuning.
      • Ablation studies?
    • Dataset(s) / data collection:
      • What data are you using or collecting?
      • Experiment design & data preprocessing / data structure
    • Results (where applicable):
      • Provide tables and figures with informative captions.
      • Report the standard error.
      • Conduct significance testing.
      • Present conclusions.
  • For psycholinguistics projects:
    • Background (The “What”)
      • The Big Picture: Briefly introduce the general area of interest.
      • Theoretical Framework: What are the core theories guiding your study?
      • Related Work: Summarize the most relevant previous findings.
    • The Research Gap (The “Why”)
      • Identify exactly what is missing in current research.
      • Clearly state your Research Question or Hypothesis.
      • Present your research question(s)
    • Methodology (The “How”)
      • Experiment design
        • Operationalize your research questions into falsifiable hypotheses
        • Define the variables of interest (dependent and independent variables)
        • Planned statistical comparisons for answering the research questions (i.e., which experimental conditions will be compared) 
      • Experiment tool
        • Eye Tracking (recommended to do it in-lab)
        • Self-paced Reading; Maze task; Mouse tracking for reading… (both in-lab and online are good)
      • Participants
        • Target demographic
          • Linguistic background (e.g., native speakers,  L2, bilingual, dyslexia, …)
          • Age range
          • Gender distribution
        • Sample size
        • if applicable:
          • How will they be recruited? In-lab or online? Via what platform?
          • Payment?
          • Experiment duration? (recommendations: less than 25 minutes for online experiment; less than 1.5 hours for in-lab experiment)
    • Materials
      • Describe the experimental stimuli you use, including areas of interest
      • Provide an example stimulus
        • If the stimuli are not in English, you must provide an English translation or explanation.
        • Use standard glossing if the grammatical structure is relevant to your point (e.g., Leipzig Glossing Rules).
      • Describe the total number of experimental items and “fillers” (distractor items)
      • Explain how items were distributed across conditions (e.g., using a Latin Square Design to ensure no participant sees the same item twice in different conditions).
      • While the above mainly applied to controlled experiment, if you are working with naturalistic reading, then please
        • Identify the origin of the texts
        • Report the total word count, number of sentences, and average sentence length etc., basic corpus statistics.
    • Method (in case you are using an already existing dataset, this can be called “Dataset(s)” and the Participants, Materials and Procedure sections might also go here; provide all information relevant to your study and provide a reference to the original paper presenting the dataset)
      • Technical set-up (hardware used and how it is arranged; software used)
      • Procedure: 
        • Describe the procedure of the experiment chronologically, including participant briefing, stimulus presentation incl. layout and randomization, etc. 
        • Describe what the participants will actually do during the experiment?
    • Data analysis 
      • Describe data preprocessing
      • Data Cleaning: Define your exclusion criteria (e.g., “participants excluded due to high blink rate” or “RTs < 80ms”).
      • Describe (planned) statistical data analysis: Bayesian or frequentist statistical paradigm? What model? Dependent and independent variables,  contrast coding, standardization of predictors, transformation of dependent variable, random effects structure, …
    • Results (where applicable)
      • Provide informative figures
        • Try to ensure that capitalization, colors, labels, and axes remain consistent across all figures, when possible.
      • Use tables for descriptive statistics of the datasets or complex model outputs (e.g., fixed effects, standard errors).

Notes & FAQs

Hardware & GPUs

  • We are unable to provide GPU resources to students.
  • Hardware limitations are no reason to fail you → we grade how you conduct your research. Along the same lines, hardware limitations are no excuse to conduct your research poorly.
  • Ways of how to mitigate hardware limitations can be discussed in the general meetings.

Programming projects / Master’s Project:

  • For programming projects involving an industry partner, the academic objectives and standards take priority over the company’s interests.
  • Programming projects must fall within the areas of cognitive applications/research or machine learning.

Eye-tracking data collections

  • You will have to collect data from 30 (Bachelor’s) to 50 (Master’s) participants. Please confirm the exact number with your supervisor. 
  • You will have to get an introduction to our lab by your supervisor or someone else in our group. Please talk to your supervisor to schedule a meeting early on as our lab is often occupied.
  • You will have to prepare advertisements to look for participants. We can share them through our channels if you send them to us.
  • We are using a booking system for participants to sign up for time slots. You can talk to your supervisor about how to set it up and use it.
  • We can only provide stationary eye-trackers that have to be used in our lab in AND, and no wearable ones.
  • In case you need lab access, please discuss this with your supervisor. In case your supervisor supports your request, send him/her your name, address, immatriculation number, and the number on the back of your UZH card to grant you access via your UZH card. 
  • Make sure you have a liability insurance (Haftpflichtversicherung)

Statistical modelling

  • Specify if you are using Linear Mixed-Effects Models (LMM) for continuous data (like RTs) or Generalized Linear Mixed-Effects Models (GLMM) for categorical data (like skip rate).
  • If using GLMMs, specify the link function (e.g., Logit for binomial data).
  • How are categorical variables coded? (e.g., Treatment contrast to compare to a baseline vs. Sum contrast to compare to a grand mean).
  • Never report an effect size in isolation. When reporting effect sizes, always pair then with measures of uncertainty, e.g., p-values (Frequentist) or Credible Intervals (Bayesian)
  • Never report a number without its unit. Always specify if values are in ms, log-ms, bits, accuracy %, etc.
  • When comparing models (e.g., to justify the necessity of including a random effect or an interaction), always report statistical significance tests.
  • Write informative but concise caption. Figure and table captions must be informative and self-contained. A reader should understand the visual without referring back to the main text.

General advice on written reports/thesis

Top