🚨 FYI, the
#EMNLP2024
conference website & 1st CfP is up! We welcome everyone to submit their papers 🤗
Note: We'll also be requesting a lot of SAC+AC+reviewing help in the coming days, so please keep some time for this 🙏
Deadline: June 15
Special Theme: Efficiency in Model…
EMNLP 2024 invites the submission of long and short papers featuring substantial, original, and unpublished research on empirical methods for Natural Language Processing. More info at:
#EMNLP2024
A small thing given all that's going on...but rcvd this kind letter w. words "permanent tenure" has been approved, & hence wanted to say a big THANKS to my students+family+mentors+collab+unc &
#nlproc
community!🙏
ps. told parents I'm not assisting, but associating w profs now😀
🥳🎉🍾 Humbled+excited abt this NSF-CAREER award, & the research it will fund! All credit/thnx goes to my students, mentors, collabr, family!
Disclaimer: I've had *several* rejections frm other NSF programs (it's important to share your rejctns & ackn challenges of the process)
🎉Excited+humbled to help lead the foundational AI efforts for equitable learning (via multimodal narrative environments) in this new Engage
@NSF
-AI Institute, & collab with our expert partners at NCSU Indiana Vanderbilt DP (+awsm
@uncnlp
team
@snigdhac25
@colinraffel
@shsriva
)🙏
Researchers from
@UNC
,
@NCState
and other universities are forming a new $20M
@NSF
Artificial Intelligence Institute for Engaged Learning. Learn how Carolina’s
@mohitban47
&
@unccs
faculty will create equitable educational experiences via advanced AI tools
Honored and humbled to receive the
@IITKanpur
Young Alumnus Award from my alma mater, which has been an amazing source of mentors+friends+memories and important foundation/values 🙏
All the credit for this award belongs to my mentors, students, collaborators, family/friends ❤️
Young Alumnus Awards were decided long time back but I couldn't make them public. Here is a post about it- perhaps my last announcement about IITK as Director.
The Young Alumnus Award (YAA) is an award that aims to recognise our alumni under the age of 40 who have distinguished…
I'll be in Toronto for
#ACL2023
, and looking forward to seeing everyone after a long time (I think this is my first in-person conference after 3.5 years)!
Happy to chat about research, or faculty+postdoc+phd positions, or simply life/hanging out 🙂
#ACL2023nlp
@aclmeeting
Life as an immigrant is already too much exhaustion+courage! Sad to see this new stress of visas...10yrs of my initial acad life (like sev others) were under constant anxiety to apply+maintain+renew multiple visa types, +also waste my short parent India visits in visa offices 1/N
Do US citizens understand how stressful it is to be an immigrant scientist in the US?
Far from family/friends, never ending visa bureaucracy & payments, constant fear of not being allowed back in the US, or being forced to leave your home/job/PhD due to some ban
The annual Phillip and Ruth Hettleman Prizes for Artistic and Scholarly Achievement have been awarded to four promising faculty members who exemplify groundbreaking and innovative research along with future career promise. Congrats to this year's winners!
🚨 Exciting batch of
#NeurIPS2023
papers/spotlights this week (11 students joining, say hi to them, some on job market🙂):
-VPGen+VPEval
-CoDi Any-to-Any
-LMs Teach Weaker Agents
-TIES-Merging
-Localization vs Editing?
-PanoGen
-Self-Chained VidQA+Loc
-Paxion
-AdaptiveContext
Honored and excited to give a keynote at
@aaclmeeting
!
I'll be talking about various challenging aspects of Multimodal Generative LLMs: unification, faithfulness, interpretability+evaluation, efficiency.
Looking forward to the conference + seeing everyone in Bali 🏖️🌊
#AACL2023
How you/we can help --> a thread of threads on useful (live-updating) info/awareness/resources/oxygen/donating/volunteering/calling/service...
#CovidIndiaHelp
Following up on this old thread: FYI if you haven't seen our updated webpage,
@UNCCS
removed the GRE requirement *permanently* for ALL future years! 🙂🎉 (+the dept continues to offer some additional app-fee waivers for inclusivity). Feel free to ping us!
Very important (& relatively easy to implement) steps we should ensure in our Grad-Admissions, for more accessibility/equitability to low-inc, first-gen, international, underrep applicants (pls fwd to your admissions chair +work on the points not already in place in your dept) 👉
Wow, it's JULY already & half the year's already over?!😮🙄 Here's to hoping the 2nd half of this crazy year will be more positive+safe+healthy for everyone --be kind!
(coincidentally, here's pics of 2 newly-born fawn in our backyard to timely represent new luck for us all)🦌🤞🙂
Results for the FEVER shared tasks are out! Check them out on Grateful to all 23 participating teams for their efforts! The results are pending a manual annotation of the responses submitted, and the final results and improved dataset will be out soon!
🌸Mentorship-For-All🌸
Free & all year round (indep of conf timelines, reg-fees, etc)! Took more time* than we hoped but finally live! Plz apply/spread word to all countries/underrep communities🙏
Thanks to the awesome team's efforts +Abhay for the logo!
Congrats to
@UNCNLP
students for their 5
@EMNLP2019
papers on cross-modal BERT/LXMERT, self-assembling modular network for multihopQA, semantic-drift reducing QG for QA data-augm, semantic IR for MRQA/FEVER, & auto-data-augm for dialogue. All preprints coming soon! 😀
#EMNLP2019
Finally an official YIPPIE KI YAY!!! 🎉🥳 And a deep first real breath after looooong anxious 4 days and even 4 YEARS....Here's to HEALING, REUNITING, REBUILDING. But for now, we are enjoying this Carrboro+Chapel Hill parade with infinite honking, plates+spoons, celebrating 😄😃
Thanks all for joining yest's mentoring session🙏 We'll be posting summary asap. Ques: we had ~110
#acl2020nlp
participants from so many places/stages🤯 but only 1 hour, so we shud set up a process to repeat these virtual sessions year-round irresp. of conferences & open-for-all?
Along with
@HannaHajishirzi
@RadaMihalcea
Heng Ji
@elgreco_winter
, I am hosting a newly-added
#acl2020nlp
mentoring session today(Wed July8) 12-1pm PDT on "Long-term career planning+Becoming a research leader: building your professional identity" (DM for zoom link). Looking fwd🙂
Folks 'at'
#acl2020nlp
, check out these 10 presentations by
@uncnlp
students+collaborators! Videos+details at (+pics attached).
We're also hiring postdocs/phds, ping us to chat (or any general topics/mentoring🙂)
cc
@snigdhac25
@shsriva
@colinraffel
1/3
Looking fwd to seeing everyone in Singapore & happy to chat abt research/jobs/life 🙂
Honored+excited to give the keynote at
#CoNLL2023
(thanks for the invite 🙏), our
@ACLmentorship
BoF session, & students presenting
#EMNLP2023
work on:
--Reasoning Chain/CoT Evaluation
...
🧵
#CoNLL2023
will feature not one, but TWO awesome keynote speakers!
On December 6th, we have Preslav Nakov !
@preslav_nakov
On December 7th, we have Mohit Bansal !
@mohitban47
The full program, including their talk abstracts is here:
see you there!
Excited to co-chair
#CoNLL2019
with
@alinev3010
(& our awesome area-chairs/organizers)! Call-for-papers is out, please mark calendars for May31 & spread the word! 😀
We will have 2 best paper awards (a special one for human language learning).
Invited speakers info coming soon!
This is the official Twitter account for the SIGNLL Conference on Computational Natural Language Learning (CoNLL) 2019, co-located w/
@emnlp2019
in Hong Kong.
Paper submission deadline: May 31, 2019
Call for papers is up at:
#NLProc
#conll2019
#emnlp2019
Had a great time speaking at
@indoml_sym
& meeting students+faculty+researchers (overall, kudos to organizers+volunteers on an excellent program)! 🙂
Talked about our long journey on Multimodal Generative LLMs:
(1) Unified/Universal Multimodal Learning for Generalizability,…
Day 1: Session 1 Foundational and Generative Models
Talk 3: Mohit Bansal Professor University of North Carolina at Chapel Hill
We’re really excited to play around with his state-of-the-art any-to-any multimodal models CoDi and CoDi-2 !
#indoml
#ml
#iitbombay
#aiml
Along with
@HannaHajishirzi
@RadaMihalcea
Heng Ji
@elgreco_winter
, I am hosting a newly-added
#acl2020nlp
mentoring session today(Wed July8) 12-1pm PDT on "Long-term career planning+Becoming a research leader: building your professional identity" (DM for zoom link). Looking fwd🙂
🚨Scalable Oversight 👉 Can LLMs generalize from EASY to HARD problems (where we struggle to label data w.r.t. cost+noise)?
Surprising result 👉 Supervision Gap Recovered is 70-100% across several ways of measuring data hardness (e.g., 3rd-grade to college-level tests)!
👇👇👇
Can LLMs generalize from easy to hard problems?
Models actually solve college test questions when trained on 3rd grade questions!
🚨New paper: “The Unreasonable Effectiveness of Easy Training Data for Hard Tasks”
🧵1/6
It was a pleasure to meet you
@Director_IITK
(thanks for your kind effort in setting up the YAA ceremony separately) & to relive old memories in my alma mater
@IITKanpur
+ interact w/ wonderful students+faculty via the
@DoRA_IITK
#AlmaKonnect
&
@CSEatIITK
mentorship session🙂🙏
It was a pleasure to host our Young Alumnus Awardee-2023, Prof. Mohit Bansal (BT/CSE/2008) today, in my office! We also felicitated him with the Young Alumnus Award-2023 today, as he could not attend the award ceremony.
- Prof. S. Ganesh, Director,
#IITKanpur
#IITKAlumni
#iitk
See a glimpse👇of how beautiful
@unc
+research triangle fall colors are 😍
Come join our awesome group of
@uncnlp
@unccs
students+staff+faculty (& great neighbors eg.
@duke_nlp
).
We are hiring at all levels (phd, postdocs, faculty); feel free to ping any of us with questions 🙏
Check out our
@uncnlp
group for your PhD applications! We have no GRE requirements + continue our application-fee waivers! Please RT+apply (we have postdoc openings too) & ping us for any questions!🙂
We strongly encourage diversity. Direct FAQ page link:
🔥Transformers' first-ever end-2-end multimodal demo was just released, leveraging LXMERT, SOTA model for visual Q&A!
Model by
@HaoTan5
,
@mohitban47
, with an impressive implementation in Transformers by
@avalmendoz
(
@UNCnlp
)
Notebook available here: 🤗
BREAKING: a new $1 million AI award will go to MIT professor Regina Barzilay, whose work uses ML to diagnose breast cancer & develop antibiotics.
More about Prof. Barzilay & the award from
@RealAAAI
@SquirrelAI_Edu
:
Yay, this really makes all the recruiting time+effort spent last 2 spring semesters feel worthwhile/well-spent 😍🙏 Really excited about these new additions! Welcome again to
@UNC
, all!
PS. NLP+ML folks applying for PostDocs/PhD this year, definitely apply to
@UNCCS
+
@UNCNLP
!!😀
2019/2020 brings a prestigious group of new faculty researchers, adding to our already accomplished and internationally recognized
@unccs
faculty team!
Welcome Samarjit Chakraborty,
@snigdhac25
, Bo Dai,
@psduggirala
,
@colinraffel
&
@shsriva
Our
@UNCNLP
group page is finally online: ! It contains several new code/model/data github repo's, student webpage links, fun group photos, colloquia info, etc. Thnx to
@AlexGuoHan
for leading the effort! (also we'll be using
@UNCNLP
handle actively now :-)
Very honored+grateful for receiving the best long paper award honorable mention at
#EACL2021
😃
Huge thanks to all my mentors+coauthors
@helfardy
,
@c_christodoulop
, Tom,
@mohitban47
Also thanks to
@eaclmeeting
pc/best paper committee & congrats to all winners/honorable mentions!
Welcome again to the research triangle, Sam+Bhuwan! Looking fwd to fun collaborations!
#NLProc
PhD applicants, apply to the awesome NLP hub of
@duke_nlp
+
@uncnlp
(w. several star new faculty incl. Sam, Bhuwan,
@snigdhac25
,
@shsriva
,
@colinraffel
) & just *10 miles* frm each other🙂
We're very excited to be starting an
#NLProc
lab at Duke in Fall 2021, and we're looking for PhD students! We'd especially like to encourage applicants from nontraditional backgrounds and communities historically underrepresented in the field. Apply here:
Welcome again,
@ColinRaffel
! Looking fwd to having u here soon 😀; & NLP folks applying for PhD this year, definitely apply to
@UNCCS
! In addn to Colin doing exciting NLP (e.g., see recent T5 paper: ), the awesome
@snigdhac25
+
@shsriva
have also joined us!
I'm starting a professorship in the CS department at UNC in fall 2020 (!!) and am hiring students! If you're interested in doing a PhD
@unccs
please get in touch. More info here:
Overflowing packed room for Triangle CS Distinguished Lecture by Chris Manning (
@chrmanning
) of Stanford NLP (
@stanfordnlp
)!!
@unccs
#NLProc
Title: Building Neural Network Models That Can Reason
Details:
PS. Video of his talk will be posted online soon!
🤯 Wow, be KIND! Sharing
#CovidIndia
posts is a tiny effort to give info to family/frnds fighting for things as basic as OXYGEN!
PS. I'll also be sharing many links these days (), feel free to unfollow OR u can help do your tiny part/ignore, "just saying"
I haven't been home in years. I stay up at night thinking of all the people I'll never see again. I'd like to have a home to go back to. All I can do is donate/RT so I'm boosting
#CovidIndia
posts that can help. If this bothers you, pls mute/unfollow. Don't send me DMs like this
Congrats to
@UNCNLP
students & collaborators for these new papers (6 ACL + 2 preprints) on lifelong learning, adversarial & interpretable multihopQA, visual relationships, VQA, paper-gener, video-QA/localization, HRI-commonsense. See u all
@ACL2019_Italy
! 😀
#ACL2019nlp
#NLProc
🚨 Important + fun collaboration on this new position paper ➡️➡️ "Rethinking Machine Unlearning for Large Language Models":
▪️ Eliminate undesirable data influence (e.g., sensitive/illegal info), while maintaining integrity of essential knowledge generation & not affecting…
[1/3] 🌟 Excited to share our latest work "Rethinking Machine Unlearning for LLMs" on arXiv! 🚀 Check it out:
📢Our mission? To cleanse LLMs of harmful, sensitive, or illegal data influences and refine their capabilities.
Congrats to
@UNCNLP
students for their 4
#AAAI2020
papers on multi-src domain adaptn bandits, QA on diverse modalities, dull response suppr, and modality-balanced visual dialogue. All preprints coming soon!🙂
FYI,
@HanGuo97
is applying for PhD programs! ()
🚨 Are LLMs good at logical + strategic reasoning? Are they better/worse than traditional solvers like Monte Carlo Tree Search (MCTS)?
➡️ **GTBench: Uncovering the Strategic Reasoning Limitations of LLMs via Game-Theoretic Evaluations**
A new framework+benchmark of 10 diverse…
🚨How good is your LLM at game theory?
Check out GTBench, to evaluate LLMs on logic+strategy games!
LLM-vs-LLM competition in 10 board+card games e.g Kuhn Poker, Tic-Tac-Toe, Breakthrough etc ➡️ LLMs overall lag on GT tasks & code-pretraining helps.
🧵
Congratulations to Mohit Bansal, a UNC
#CompSci
assistant professor, on receiving the prestigious Young Investigator Program Award from the U.S. Army Research Office! The award provides $360,000 to fund his research.
Congrats to
@peterbhase
for 'Outstanding Area Chair' award at
#ACL2023nlp
🙂
On a related note, he will be on the faculty market this year, so definitely look out for him 👉
#ProudAdvisor
And we continue the distractions. My latest for
@techreview
covers a paper from
@HaoTan5
&
@mohitban47
that cracks how to train neural networks on language & images through mostly unsupervised learning. A promising direction for developing common sense AI.
🎉🎉Congrats to
@jayleicn
+
@linjiefun
on the
#CVPR2021
Best Paper Nomination for "Less is More: ClipBERT for Video-and-Language Learning via Sparse Sampling": efficient/democratized + sota end2end video&lang learning!
w/
@luowei_zhou
@zhegan4
tlberg jjliu
🚨 We are hiring assistant+associate level faculty for the
@unccs
family, including areas broadly in DS/ML/AI+x. Please feel free to apply & reach out for any questions 🙏
We're hiring tenure-track faculty here at
@unccs
! Areas include: Security, Data Science, Vision, Systems, Robotics, CompBio & AR/VR. Join a dept renowned for research collaboration and impact.
@unccollege
@UNCResearch
@UNC
Congrats Jie
@jayleicn
on being awarded the
@AdobeResearch
Fellowship!! 🥳 🎉
#ProudAdvisor
PS. Jie works on diverse aspects of vision+language research (esp. videos) & will be on the job market next year 🙂
(cc tlberg;
@uncnlp
)
Reminder to self/others in current isolation situation: We are spending *way* more time online on screens than usual (due to all meetings --> virtual), so plz take v good care of ur eyes/health via regular screen/exercise breaks; u might even spot a backyard frnd in such breaks🙂
We're excited to announce the 2019 winners of the Microsoft Research Ada Lovelace Fellowship and PhD Fellowship! Explore inspiring projects from talented PhD students in computing-related fields:
🎉🥳😍 Congratulations to
@jayleicn
+
@linjiefun
+the whole team on the
@CVPR
Best Student Paper Award Honorable Mention (and to all the other winners/papers' authors)!!
🚨 "Evaluating Very Long-Term Conversational Memory of LLM Agents"
➡️ we evaluate 'very long + multimodal conversational memory' of LLMs on our LoCoMo dataset via 3 tasks: QA, Event Summarization, Multimodal Dialog Generation.
➡️ we show that LLMs (incl. long-context models)…
Can LLMs keep track of very long conversations?
We evaluate 'conversational memory' of LLMs via 3 tasks on our dataset of multi-session multimodal dialogs --> LLMs struggle to remember, reason over history, draw long-range temporal/causal connections
🧵
Please RT: I am looking for a POSTDOC in NLP/ML for
@UNCNLP
!! Broad focus, freedom to create research agenda, advise students, collab w/ other AI faculty.
More info:
Ping me or happy to chat at
@EMNLP2018
!
Here are some new upcoming
#NAACL2019
,
#CVPR2019
,
#ICRA2019
papers from our students & collaborators on {Navigation, Self-Learned-MTL, Summ-Eval, EQA, Attribute-Instruction-Following} + RANK-1 in popular Room-to-Room Vision-Lang-Navigation Leaderboard ()!
🚨
#TMLR
paper on exposing+addressing how large VL models show significant **Cross-Task Inconsistency** for the same image across different tasks! 😲
E.g. the model describes [captioning task] the image as a person 'skiing' BUT answers 'snowboarding' when asked [vqa task] the…
Are unified VL models consistent across predictions for different tasks on the same image?
Thrilled to share our
@TMLRorg
paper where we find that VL models show significant cross-task inconsistency in their predictions for the same image across tasks.
🧵
🚨SouthNLP🚨
Looking forward to this exciting regional symposium involving 10+ universities (& everyone welcome)! Thanks to Jinho+Emory+organizers for their effort on 1st version; hope to see many of you in Atlanta + check the CfP and submit your papers 👉
We are delighted to announce that the First South NLP Symposium will take place on March 29, 2024, at Emory University:
Here are the details regarding the call for papers:
We look forward to seeing you at the symposium.
"Vokens" = Visually-grounded-tokens (contextual) to imprv lang-pretraining & engl NLU tasks (imp divergence/grounding ratio issues, extrapolates frm small dataset)!
pdf:
Full code:
➡️Hao is on job market🙂:
*Vokenization*: a visually-supervised language model attempt in our
#emnlp2020
paper: (w.
@mohitban47
)
To improve language pre-training, we extrapolate multimodal alignments to lang-only data by contextually mapping tokens to related images ("vokens") 1/4
How to efficiently & flexibly inject diverse new modalities (RGB+audio/3D/depth/flow/...) to improve Video Reasoning?
Introducing ☕️ CREMA, a substantially efficient (~3% trainable params) & modular fusion framework (w/o complex architecture changes)
🧵
🎉🎉 Yay congrats Shiyue on being awarded the
@Bloomberg
PhD Fellowship, for her impactful work on low-resource NLG (eg. Cherokee revitalization) & summarization!
#ProudAdvisor
🙂
PS. She'll be on job market in 1 year and is awesome, definitely get her ->
I am very honored to receive a
@Bloomberg
Data Science Ph.D. Fellowship!
Many thanks to my advisor
@mohitban47
for all the support and guidance I received. I’d also like to thank all my labmates
@uncnlp
, collaborators, and mentors for their help all along!
It was great chatting with
@samcharrington
on this
@twimlai
podcast (in April), on various different aspects of Multimodality & Vision-Language LLMs, including Unification, Efficiency, Evaluation, and Explainability...(PS. don't mind my long hair 😄) -- thanks for having me!
Today we’re joined by Mohit Bansal (
@mohitban47
), Professor at
@unccs
, to discuss his work on the unification of language and vision models, the advancement he’s seen in multimodal models, challenges of evaluation in generative AI, and much more!
🎧
🎉🎉BIG congrats to
@ZinengTang
for the amazing achievement of being selected as Winner (out of 4 in North America) of the 2023 CRA Outstanding Undergraduate Researcher Award!
#ProudAdvisor
🙂
🚨 Zineng is applying for a PhD this year 👉
@CRAtweets
@unccs
Thanks everyone for coming to
@CoNLL2019
! It was great to see so many strong presentations! Thanks to my awesome co-chair
@alinev3010
+ our invited speakers
@chrmanning
&
@vigliocco_g
+ amazing orgnz team (incl
@emnlp2019
org)! Looking fwd to CoNLL2020 in Dominican Republic! 🙏🙂
Big congrats to the winners of the
#conll2019
paper awards: 1) Best Paper: by Gonen, Kementchedjhieva, Goldberg; 2) Best Paper for Research Inspired by Human Language Learning: by Rabinovich, Watson, Beekhuizen, Stevenson.
Submit to our
#neurips2020
"HAMLETS: Human And Machine in-the-Loop Evaluation and Learning Strategies" workshop!
Deadline Oct9 -->
And check out the (growing) list of exciting speakers: 🙂
Call for papers for the
#neurips2020
HAMLETS workshop is live along with a great list of invited speakers including
@jennwvaughan
,
@ajratner
,
@EmmaBrunskill
, Sanjoy Dasgupta, Dan Weld, Kristen Grauman, and Finale Doshi-Velez. More to come.
Submit!!!!
Wow, no time to catch one's breath between such news items ☹️ I was on F1 too. Making students choose between health safety vs visa safety?!! 🤯 (but pls don't panic, we'll work v hard with univs to extend in-progress hybrid solutions & distance-safe inperson indep-study courses)
This ICE action against students on F-1 visas is incredibly damaging in many ways. It hurts U.S. competitiveness, damages our world-class universities, and, if enacted, will disrupt the scholarly study of millions of the brightest students from around the world.
Several tenure + teaching track faculty positions at
@unccs
! Please apply/share at + & feel free to ping us with any questions that we can help with 🙏
Join the tradition of teaching excellence, innovative research, and collaboration at
@unccs
. We are HIRING for multiple faculty positions. Areas include: security, data science, vision, systems, robotics, compbio, & AR/VR. Apply:
@unccollege
@UNCResearch
Today, our old backyard friend came to visit on the front porch and even peaked into our window to say hello😀Also brought the family along!! Served as a good reminder to regularly take those important screen/exercise breaks for eyes/health and appreciate nature!
Reminder to self/others in current isolation situation: We are spending *way* more time online on screens than usual (due to all meetings --> virtual), so plz take v good care of ur eyes/health via regular screen/exercise breaks; u might even spot a backyard frnd in such breaks🙂
🎬🎥Can LLMs create a detailed “video plan” that describes entities/background/layouts/consistency-groups for long video generation?
VideoDirectorGPT: Consistent Multi-scene Video Generation via LLM-Guided Planning
@AbhayZala7
@jmin__cho
@mohitban47
🧵
Presenting our new V+L pretraining work: “Unifying Vision-and-Language Tasks via Text Generation”,
a single unified generative framework (VL-T5 / VL-BART) for diverse multimodal tasks!
Arxiv:
Work done w/
@jayleicn
@HaoTan5
@mohitban47
(
@uncnlp
)
🧵1/n
Our
@RealAAAI
paper "Multi-Source Domain Adaptation for Text Classification via DistanceNet-Bandits" (w/
@ramakanth1729
,
@mohitban47
): we analyze distance measures for NLP domain adaptn; DistanceNet using a mixture of distances; & bandit that dynamically switches btw src domains
Inspiring article on
@UNC
univ page abt our awsm undergrad
@SwetaKarlekar
! Proud of her impact: CRA Outstanding UG Award fr NAACL+EMNLP+WiNLP papers on Alzheimers+SexualHarassment+
#MeToo
, & AI+ethics work at FB,Google,Disney,Yelp! Best wishes fr
@Facebook
job, & Go class
#UNC20
🙂
Sweta Karlekar had some extraordinary experiences at
#UNC
, but she'll miss little things — like studying at Davis Library until 2 a.m. — most of all. As she heads to work at
@Facebook
, the new
#UNCgrad
urges her classmates not to give up on their dreams
🚨 EnvGen --> LLM iteratively decides+generates visual/embodied environments that are most effective to let the RL game agent play+learn in
--> so as to automatically target/focus on progressively/adaptively improving the agent's weaker skills
--> and hence very efficient…
Can we adaptively generate training environments with LLMs to help small embodied RL game agents learn useful skills that they are weak at? 🤔
👉 Check out EnvGen, an effective+efficient framework in which an LLM progressively generates and adapts training environments based on…
Can we teach multiple skills to a text-to-image (T2I) model (w/o expensive annotations), while minimizing knowledge conflicts between skills? 🤔
👉SELMA improves T2I models by fine-tuning on automatically generated multi-skill image-text datasets, with skill-specific LoRA expert…
🚨Exciting new work on reliable Text-to-Image Evaluation led by
@jmin__cho
(after his prev works on DALL-Eval, VPEval, LayoutBench, etc.) with a great group of collaborators🙂 based on Davidsonian Scene Graphs for atomic + unique + semantic coverage + valid dependency questions👇
🚨New T2I Evaluation!🚨
We introduce Davidsonian Scene Graph (DSG) for reliable T2I evaluation with questions that:
- are atomic and unique
- cover full text prompt semantics (w/o hallucination)
- and have valid consistencies
@GoogleAI
@uncnlp
@uwnlp
🧵
Folks applying for O1/greencard style visas, check your updated citations because the jumps are large (& no email notification)! Thanks again
@annargrs
+
@earnmyturns
: this also really helps intl immigrants because O1/GC type apps (unfortunately) put high weight on citations...1/2
This long-standing issue seems to have been resolved. Thanks to
@annargrs
for raising this issue,
@earnmyturns
for raising this issue internally, and also to Google Scholar for the fix!
🤩 YAY we are super excited
@EliasEskin
to have you join us as a postdoc!! 🥳
Welcome to the MURGe-Lab +
@uncnlp
+
@unccs
family & the beautiful Chapel Hill + Research Triangle area!! Looking forward to the many fun collaborations 🙂
After an amazing 5 years as a PhD student at
@jhuclsp
, I'm incredibly excited to announce that I'm joining
@mohitban47
's lab
@uncnlp
and
@unccs
as a Postdoctoral Research Associate!
Looking forward to many exciting collaborations with the amazing students and faculty there
1/2
Diagrammer ➡️ Diagram Programmer+Generator!
-Plans for fine-grained, dense layouts+relationships/connections for open-domain/platform diagram gen (on which sota T2I models struggle).
-Visualization of complex concepts & human-in-the-loop diagram layout editing via ppt/inkscape.
For me, it was Lillian Lee & Claire Cardie (
@clairecardie
), who were my co-mentors for an undergrad summer fellowship from India to Cornell. As I told them recently (was lucky enough to host both of them for colloquia at UNC), my interest in NLP started strongly because of them!!
In honor of
#InternationalWomansDay
tomorrow, I'd like to start a thread about women who have inspired us, taught us, and mentored us in our careers. Please comment with someone who has inspired you! (cont)
Very important (& relatively easy to implement) steps we should ensure in our Grad-Admissions, for more accessibility/equitability to low-inc, first-gen, international, underrep applicants (pls fwd to your admissions chair +work on the points not already in place in your dept) 👉
Hear Hear!! Please read again, reviewers (esp. omnipresent reviewer
#2
🙂), & also ACs+SACs when writing/revising/discussing reviews in upcoming weeks.
And a big thanks again to
#emnlp2020
team and
#nlproc
community for leading the way in helping inculcate these important values
Seems like a lot of people are working on their EMNLP reviews at the moment! Please remember that the following are NOT valid primary reasons for rejection, and the authors will be able to appeal to these guidelines in the rebuttal:
🥳 🎉 Big congrats
@peterbhase
on being awarded the
@GoogleAI
PhD Fellowship in
#NLProc
, for extensive work on diverse aspects of explainability!!
#ProudAdvisor
🙂
(and congrats to all the other winners too!)
I am honored to receive a
@GoogleAI
PhD Fellowship this year! This could have never happened without the support of my advisor
@mohitban47
and
@uncnlp
, great co-authors, and many prior mentors. Looking forward to more work on NLP and AI Safety
Congrats to
@PeterBHase
's for 1st PhD work at
#ACL2020nlp
on "Evaluating Explainable-AI".
Important take-aways from comprehensive fwd+counterfactual simuln tests on text+tabular data for 5 XAI methods(inc. prev 'prototype' work w Cynthia Rudin) via *many* in-person user studies🙂
My first PhD paper!😀 (at
#acl2020nlp
, w.
@mohitban47
@uncnlp
)
"Evaluating Explainable AI: Which Algorithmic Explanations Help Users Predict Model Behavior?"
We measure how 5 explanation methods (LIME, Anchor, Prototype, Decision Boundary, Composite) improve simulatability...1/5
🤩A BIG WELCOME to Dr.
@jaeh0ng_yoon
(), who is joining
@uncnlp
@unccs
as a Postdoc this fall from KAIST!
Looking forward to the many exciting collaborations with our students+faculty, Jaehong! 🙂
😍I'm super excited to announce my next journey! After a great time at KAIST, I'll be working as a Postdoctoral Research Associate at UNC Chapel Hill (
@UNC
) this fall, working with Prof. Mohit Bansal (
@mohitban47
) and faculty+students in the awesome
@uncnlp
and
@unccs
groups!
1/3
Very important to explicitly list these out for inculcating their importance in both reviewers+authors, esp w. drastic conference+reviewer pool size increase these days. So everyone please read this v carefully & spread the word! Great job
#NLProc
community, in leading the way!🙂
New blog post on reviewing policies!
Special focus on spurious reviews: this time no paper should be rejected primarily for not beating SOTA, for non-English work, for being a resource paper, etc. /1
Join our next timely
@ACLmentorship
session on HCI+NLP, with experts
@Diyi_Yang
@TongshuangWu
@XuWangHCI
!
We'll start w/ short talks abt their work & then Q&A/advice (remember to send us questions; zoom+slack👉)
PS. Also help sign up as future mentors! 🙂
Very excited to host Prof. Kathy McKeown of
@ColumbiaCompSci
for her Triangle CS Distinguished Lecture this Mon Feb3 (4pm SN011; recorded) -- join us & ping me for any meeting requests!😀
"Where Natural Language Processing Meets Societal Needs"
Details:
Join us for the latest Triangle Distinguished Lecture from
@Columbia
's Prof. Kathleen R. McKeown (Henry & Gertrude Rothschild Professor of
@ColumbiaCompSci
and the Founding Director of
@DSI_Columbia
.) on Feb. 3 (Monday) at 4 PM, SN011.
For appts. please contact host
@mohitban47
Very proud of our
@UNCNLP
undergraduate researchers
@SwetaKarlekar
&
@HanGuo97
for this amazing achievement of being selected Runner-Up & Finalist for the 2020 CRA Outstanding Undergraduate Researcher Award!!😀
#NLProc
@CRAtweets
🤯🙀 Saw this late but important to repeat: please NEVER aim for 90hrs/week work!! Will harm health & cause burnout/anti-productivity. It should anyway be impossible to find even close to 90 hrs/week of non-sleep/exercise/food/relax time (9-10hrs/day goes in sleep itself)?!! 👇👇
Re-posting this due to popular demand. Do not aim to work 90 hours a week. I repeat, do not aim to work 90 hours a week.
#AI
#ML
#dlearn
#NLProc
(if you've not seen it yet, check out this thread: )