Project Interest Survey (M1)

LING 282/482, Fall 2026

Due: Friday, September 18, 2026

Submit via: Google Form (link on Blackboard)

This survey helps me understand your interests and background so I can provide better guidance on your term project. Working in a group is highly encouraged but not required, and your responses to this survey are what I use to announce suggested groupings after this milestone — so it’s worth answering carefully.

Question 1: Project Ideas

Question type: Long answer (paragraph)

Question text: What ideas do you have for potential topics for your term project?

Description/Helper text: It’s early in the course, so you may not have an exact research question or methodology in mind. However, you can still identify topic areas or personal interests. For instance:

  • Are there specific aspects of language you’re curious about? (e.g. syntax, morphology, semantics, pragmatics, phonetics/phonology, historical change, language acquisition, sociolinguistic variation)
  • Are there specific languages or varieties you’d like to work on? (Including low-resource languages, dialects, code-switched data, or historical stages of a language)
  • Are there questions about models rather than about language that interest you? (e.g. what do neural models represent internally? how do they generalize? how do they compare to human learners?)
  • Any model types or methods you’re already curious about? (e.g. word embeddings, RNNs, Transformers, pre-trained language models, speech models)

Remember that this project is scientifically oriented: the goal is to answer a question about language or about language models, not to build a system that performs a task well. Please answer in a paragraph or two, being as specific as possible. If you already have a research question in mind, feel free to share it — but it’s totally fine if you’re still exploring!

Question 2: Languages of Interest

Question type: Long answer (paragraph)

Question text: Are there particular languages or language varieties you’d be interested in studying for your project?

Description/Helper text: This doesn’t have to be a language you speak — interest is what matters here. Low-resource languages, dialects and regional varieties, code-switched data, sign languages, and historical stages of a language are all fair game, and all are studied far less than English.

That said, it is genuinely useful to have a native or fluent speaker of the language on the team: someone who can sanity-check model output, judge whether an apparent error is really an error, and tell whether a corpus is what it claims to be. So if you do speak a language you’re interested in working on, mention that — and if you don’t, it’s one good reason to look for a groupmate who does.

If you have no particular preference, or expect to work on English, just say so.

Question 3: Linguistics Background

Question type: Long answer (paragraph)

Question text: What background do you have in linguistics, and in what subfields?

Description/Helper text: This course is cross-listed and students arrive from very different starting points. Some of you have a linguistics degree and little programming experience; some have written a lot of code and have never taken a linguistics course. Both are fine, and neither is a disadvantage for the project — but it helps me suggest groupings with complementary backgrounds, and it tells me what kind of guidance will be useful to you.

Mention relevant coursework, research, or self-study. If you have none, just say so.

Question 4: Data Access and Availability

Question type: Multiple choice

Question text: Do you already have access to language data you’d like to use for your project?

Options:

  • Yes, I have a specific corpus or dataset in mind and can access it
  • I have data access through my research/work but need to check permissions
  • No, but I know what kind of data I want to find
  • No, I’m open to using any publicly available corpus
  • Not sure yet

Description/Helper text: Data access is often the biggest blocker for projects. Knowing your situation early helps me guide you toward feasible options. Some linguistic resources are licensed rather than open (e.g. many LDC corpora), and some are open but far messier than they appear. If you selected “Yes” or have data access through work/research, please mention the dataset briefly in Question 1 (Project Ideas) above.

Question 5: Learning Goals

Question type: Checkboxes (select all that apply)

Question text: What specific topics or skills from this course are you most excited to learn?

Options:

  • Word vectors and distributional semantics
  • Language modeling
  • Recurrent neural networks (RNNs, LSTMs)
  • Transformers and attention
  • Pre-training and fine-tuning
  • Tokenization and subword models
  • Multilingual language models
  • Large language models
  • Speech data, acoustics, and speech models
  • Probing and interpreting what models represent
  • Working with computing clusters and GPU resources
  • Reading and implementing research papers
  • Other (please specify in Additional Information below)

Description/Helper text: Select all that interest you. This helps me understand what you’d like to get out of the course and can help match you with appropriate project topics and groupmates.

Question 6: Group Preferences

Question type: Short answer

Question text: Are there any classmates you’d prefer to work with on your project?

Description/Helper text: Working in a group is highly encouraged but not required, and group formation is ultimately up to you. I’ll use these responses, along with shared topic interests and complementary backgrounds, to announce suggested groupings after this milestone.

Note that I expect the same quality of work from an individual project as from a group project — the scope a group can take on is simply larger. If you’re leaning toward working solo, that’s allowed, but go in with that expectation in mind.

If you’d like to work with specific classmates, list them here. If you have no preferences yet, or plan to work alone, just write “No preference” or “Solo.”

Question 7: Python Programming

Question type: Linear scale (1-5)

Question text: Rate your comfort level with Python programming

Scale:

  • 1 = Beginner (limited experience)
  • 2 = Some experience (can write basic scripts)
  • 3 = Comfortable (can write programs that read, process, and write data files)
  • 4 = Proficient (comfortable with NumPy and similar libraries, can read and modify research code)
  • 5 = Expert (regularly write research code, familiar with deep learning frameworks)

Question 8: Math Background

Question type: Linear scale (1-5)

Question text: Rate your comfort level with the math underlying deep learning (linear algebra and calculus)

Scale:

  • 1 = Beginner (haven’t studied this, or it’s been a long time)
  • 2 = Some familiarity (took the prerequisite calculus, rusty on the details)
  • 3 = Comfortable (comfortable with derivatives and with vectors and matrices)
  • 4 = Proficient (comfortable with the chain rule, matrix multiplication, and gradients)
  • 5 = Expert (have coursework or research experience in linear algebra and multivariable calculus)

Description/Helper text: This is here so I know how much time to spend on the math, not to screen anyone. The course prerequisite is one semester of calculus, and we build up the linear algebra from the start.

Question 9: Command-Line and Computing Clusters

Question type: Linear scale (1-5)

Question text: Rate your comfort level with command-line interfaces and computing clusters

Scale:

  • 1 = Beginner (rarely or never used command line)
  • 2 = Some familiarity (can navigate directories, run basic commands)
  • 3 = Comfortable (can use SSH, run scripts remotely)
  • 4 = Proficient (comfortable with job schedulers, environment management)
  • 5 = Expert (regularly use HPC resources for research)

Description/Helper text: Don’t worry if you’re not comfortable with this yet — we’ll have a hands-on session to get everyone set up on the BlueHive cluster in class on Sep 28.

Question 10: Additional Information (Optional)

Question type: Long answer (paragraph)

Question text: Is there anything else I should know that would help me support your learning this semester?

Description/Helper text: This could include:

  • Specific learning goals for the course
  • Constraints on your time or availability
  • Preferred communication style
  • Accessibility needs
  • Anything else you think would be helpful