Kanishk Verma
@vermakanishkk
Followers
50
Following
369
Media
1
Statuses
35
PhD student, @irishresearch @google Scholar, at @adaptcentre @antibullyingcen @dcu
Dublin City, Ireland
Joined October 2017
I have a recommended reading list for Artificial Intelligence, and it hasn't changed since 2019. I give this list to my grad students, but all of the articles are broadly accessible if you're interested. Very short 🧵.
44
383
3K
Interesting work, let's collaborate on social media language data.
A Weekend Project that Saw the Light of Day: Implemented a variant based on universal transformers paper: https://t.co/6gREDa58ii
0
1
0
Join me at #LRECCOLING2024 today as I present our Research Poster ! 🕓 9am - 10:40am 📍Poster Area II Topic: Beyond Binary: Towards Embracing Complexities in Cyberbullying Detection and Intervention - A Position Paper Paper: https://t.co/5rMWEmOf8F
#NLProc #Turin #Italy
aclanthology.org
Kanishk Verma, Kolawole John Adebayo, Joachim Wagner, Megan Reynolds, Rebecca Umbach, Tijana Milosevic, Brian Davis. Proceedings of the 2024 Joint International Conference on Computational Linguist...
0
0
4
Delighted to receive the Best Student Contribution Award at the @AdaptCentre Annual Scientific Conference #AASC2024 . I'm immensely grateful for the guidance and continued support of my mentors and researchers at Adapt ! #StudentAward #PhDJourney
ADAPT held our third Annual Scientific Conference in celebration of the research carried out in the centre this week in @DCU. Congratulations to our #AASC2024 winners: 🏆@simon_mille for Best Overall Contribution Award (accepted by Prof. @anyabelz) 🏆Jamsheela Nazeer for Best
1
0
4
🚨Two New Papers🚨I've been building LLMs to persuade humans, now it's time to #persuade #LLMs ! (1) How to #persuade LLMs to #jailbreak them with a success rate of 92%, in a single-turn dialogue Project: https://t.co/uLNmaLUrjV Paper: https://t.co/gbxhQKkcyH (2) How to
🚨 [New Paper] If you're involved in AI safety or jailbreaking, you don't want to miss this: Techniques from human communication now effectively breach aligned LLMs (Llama-2 Chat, GPT-3.5, GPT-4) with over 92% attack success rate. 👇🧵(1/7 - page link: https://t.co/FJhgRzHxN5)
10
60
307
Really enjoyed sharing the results of our research on #AI #children's rights @ Adoprivacy today! @vermakanishkk @AdaptCentre @AntiBullyingCen
0
2
8
Prof Brian Davis @DCU discusses harnessing the power of pre-trained language models (#PLMs) for annotation efficiency & the importance of addressing potential biases & limitations present in the models at #ADAPTSciCon2023 #LLMs #ASC2023
0
1
2
Oops haven't tweeted too much recently; I'm mostly watching with interest the open source LLM ecosystem experiencing early signs of a cambrian explosion. Roughly speaking the story as of now: 1. Pretraining LLM base models remains very expensive. Think: supercomputer + months.
147
927
6K
"In a recent study, Dr Milosevic asked teenagers about their views on the effectiveness of such AI-driven cyber bullying regulation on social media platforms." @AntiBullyingCen @EliteFellowship
independent.ie
This article is on the pros and cons of artificial intelligence (AI) and these two opening paragraphs have been written by me, ChatGPT.
0
2
3
As part of @DCU Ethics & Privacy Week, FUJO' @CullotyEileen spoke about the limitations of platforms’ transparency reporting for today’s online safety panel with @TiMilosevic and @vermakanishkk
0
6
16
Final webinar of the DCU Ethics and Privacy Week '23 today at 1pm! With @CullotyEileen @TiMilosevic @vermakanishkk we'll discuss Digital Safety: AI, Social Media, Bias and Harm. And ofc the new Online Safety and Media Regulation Act 2022! Link 👇
eventbrite.ie
DCU Ethics and Privacy Week 2023
0
2
4
Looking forward to be on the panel for @DCU Ethics and Privacy Week 2023, Digital Safety: AI, Social media, Online Harm on Thursday, 2nd Feb. Check the link below for registration details, https://t.co/DTCiaKek5A
0
1
3
I wrote this piece on masculinity, the manosphere and mental health because so many boys are saying that Tate is good for mental health and self-improvement advice. He’s not. https://t.co/qkCDwx55E8
irishtimes.com
Wellness rhetoric functions as a Trojan horse to smuggle in a more sinister agenda
21
84
245
Recently, one of our researchers @debbieging spoke to RTE about the concerns of Andrew Tate's influence on young men and boys. Listen to the 10 minute clip here👇 https://t.co/tgtxN1yYsn
0
1
8
Co-Designing With Children: A Rights-Based Approach to Fighting Bullying Join @TiMilosevic from @AntiBullyingCen & @AdaptCentre at @DCU_CER's "Sharing Knowledge" online seminar series on 26 Jan at 12.15. Register now! #EngagedResearch #DiscussAI
https://t.co/FFBWxyotXa
0
3
3
Children raise concerns around privacy and effectiveness of interventions, while broadly (in theory) welcoming AI, provided there is transparency and opt-in
First study in Ireland to consult teens 12-17 yrs of age on AI to Address Cyberbullying. Research finds Teen Girls Fear Being Perceived as Being 'Weak' for Using AI Interventions, research by @TiMilosevic @AntiBullyingCen read more in @businessposthq:
0
5
7
ABC were delighted to host @AphraK to our @gegameproject @EUErasmusPlus launch of #GenTOPIA video game for #gender equality. A special thank you to the young people from @Sphere17RYS @CTYI @myIADT @clonturkcc who attended and played the game: https://t.co/Jt8fHdwQE2
@dereklaffan
0
5
11
Declining user reporting of #cyberbullying (why?) and rising AI detection of #cyberbullying (what?). #AoIR2022 @TiMilosevic #KanishkVerma
3
3
28
Superfast #stablediffusion with JAX on TPUs & GPUs 🧨diffusers 0.5.0 now supports JAX! What does that mean? Many cool things, including that you can generate 8 images in ~8s on Colab using TPU 🤯 https://t.co/F9BujBtjjX Learn how to use JAX with diffusers
huggingface.co
8
112
554