Joe Stacey
@_joestacey_
Followers
2K
Following
6K
Media
62
Statuses
595
NLP PhD student at @ImperialCollege and @Apple AI/ML Scholar. Soon to be at @SheffieldNLP!
Joined September 2020
We have a new paper up on arXiv! 🥳🪇. The paper tries to improve the robustness of closed-source LLMs fine-tuned on NLI, assuming a realistic training budget of 10k training examples. Here's a 60 second rundown of what we found!
3
16
77
I’m so excited to be on my way to the US!! I’m doing a little end of PhD tour 😍😍. Thanks so much to all the US unis that are having me for visits, I honestly can’t wait 🙂.
2
2
30
I would love to see a quality score for each reviewer in ARR alongside their reviews. Eg expanding the ‘great review’ process now, each review could be rated 1-5 by ACs and then we could see the avg rating of each reviewer next to any of their reviews. It would be so helpful to.
4
0
38
RT @DrDominicNg: Microsoft claims their new AI framework diagnoses 4x better than doctors. I'm a medical doctor and I actually read the pa….
0
1K
0
The closer to my thesis deadline, the more I'm wasting my time on Twitter 🤦♂️🤦♂️🤦♂️🤦♂️.
3
0
18
If you're interested in interpretable methods, here's a little blog post for some of the work me and my amazing collaborators were doing in this area:.
0
0
2
I was re-reading this article in Nature by Cynthia Rudin about the need for interpretable models. The paper is so nice, and it seriously deserves the thousands of citations it has. 6 years later and I still think it's excellent
1
1
10
The UK heatwave at the moment really is no joke. 32 degrees yesterday and it's totally unbearable. Grey skies and rain again please 🙏 .
0
0
6
One day after the ARR review deadline, and surprisingly few other reviews done for the submissions I'm reviewing. I know other reviewers will be much more senior than me with way more responsibilities, but it must be so stressful for the organisers 🙂.
1
0
21
RT @oanacamb: 🥳Thrilled to announce that I started as an Assistant Professor @imperialcollege! Grateful to all my mentors and collaborators….
0
24
0
For anyone wondering I'm stopping in a town called Meridian in Mississippi. It seemed like a really cool place to visit 😍.
0
0
2
I know it's weird, but I'm totally obsessed with Amtrak long distance train journeys. I've planned one more for summer, from Washington DC to New Orleans, which is 27ish hours plus a 24 hour stop in Mississippi. I literally can't imagine anything more exciting in this world :).
1
0
9
As someone who didn’t drink caffeinated drinks, iced coffee on a hot day is a real gateway drug.
1
0
9
I'm planning a little US trip for summer, and wow did I get my US geography wrong. My trip starts: Utah > Kentucky > Texas. I was confused about why my flights were so long, so I looked at an actual map, and it turns out Kentucky is definitely not between Utah and Texas 😂. I.
1
0
8
Super excited to have a cool post-doc lined up for after my PhD, working on Uncertainty Quantification with @NafiseSadat at @SheffieldNLP, and @benbenhh and @inuikentaro from RIKEN, Japan. Could not be happier 🙂.
10
3
87
My twitter followers are so awesome, everyone is so nice and supportive all the time. Really appreciate all of you 🙂.
1
0
24
If anyone has done any work improving the robustness of NLI models and we didn't cite you in our appendix, please share a link to your work - I would love to include it. Our appendix related work on NLI robustness is a bit of a monster 😅🧌.
This work was really fun and a great last paper for my PhD. Check it out 🙂. P.S. if you know about a paper improving NLI model robustness not already in our related work appendix, I would love to hear about it🥰.
2
3
10