/Tech31d ago

Graham Neubig at CMU's Language Technologies Institute says most projects originate from collaborator work rather than principal investigators and recommends tying research to real applications via startups

AI Judge changed title after evaluation, original title: "Graham Neubig, associate professor at CMU's Language Technologies Institute who leads NeuLab and co-founded OpenHands, described core methods for developing research ideas, noting most pursued at CMU arise from collaborators rather than principal investigators alone"

Peng Qi agrees external researcher conversations aid idea generation.

72031914016.2K

#89

Original post

Graham Neubig#89

Graham Neubig@gneubig#89inTech

Great question about how to come up with research ideas!

I think that different people have different ways, but here are some of mine:

1. Working with excellent people -- most of the ideas we tackle at CMU originally come from the people I work with, not me 2. Thinking about impact -- I always ask my collaborators "who is your target audience, and what results will they need to see to be convinced that this is promising" 3. Staying close to real-world use -- as you know, I have a startup and many of the ideas that we have for research come from practical problems we need to solve there too

(and sorry about the late reply, I needed to think about this before responding)

Christopher@communicating

@gneubig @yueqi_song Honest question. How do you select the specific topics within your research area you and your team go deep on vs others?

I feel you consistently put out papers (with artifacts - thank you) that are not only must reads but present ideas that are implementation ready.

6:55 AM · May 17, 2026 · 13.9K Views

Sentiment

Sentiment building, check back later.

Cluster Engagement

Digg Deeper

No Digg Deeper questions have been answered for this story yet.

GITHUB.COMVia

Posts from X

Most Activity

VIEWS2.2KBOOKMARKS18LIKES8REPLIES2RETWEETS2

Toni Kukurin@tkukurin

@gneubig @communicating @yueqi_song cf. also @lateinteraction's similarly terrific reflection in blogpost form https://github.com/okhat/blog/blob/main/2024.09.impact.md

Graham Neubig@gneubig

Great question about how to come up with research ideas!

I think that different people have different ways, but here are some of mine:

(and sorry about the late reply, I needed to think about this before responding)

31d2.2K818

Peng Qi@qi2peng2

@gneubig @communicating @yueqi_song +1 to all of these.

One thing I personally find helpful (rel. to #1) -- talk with excellent people (esp non-collabs) often, even if you don't know what to work on yet. Exchanging ideas have always paid off well for ideation, just like error analysis does for experiments.

Graham Neubig@gneubig

Great question about how to come up with research ideas!

I think that different people have different ways, but here are some of mine:

(and sorry about the late reply, I needed to think about this before responding)

30d6720