Updates as promised, this summer I graduated from
@Stanford
with a Masters in Symbolic Systems (CS, Linguistics, Psych and Phil) and a novice degree in life.
The past two years have been the most enriching and I'm truly grateful for every up and down!!
@stanfordsymsys
🌲💕🎓
Was genuinely surprised to not only see Grant while having dinner but also to be able to talk to him and discuss which topics could be added to the linear algebra series. Ah, this is going to take a while to get over.
What better day to announce this: I am joining
@nvidia
as an ML Solutions Architect!
I will be working with Cloud Providers to adopt and adapt Nvidia HW + SW stack for their needs. Absolutely stoked! 🥳🎉
This week I started Student Researcher position at
@Google
with the Translate team. Could not ask for a better host
@adisid01
and a more wholesome team. 3 years and 14 interviews later 💙❤️💛💚
In my junior year of undergrad, I worked with a reputed professor from a very esteemed university of India. There would be talks about my alcohol preferences in the 8 months of internship, but I took it in light humour. After my project, the prof wanted me to work on a project..
Walter Lasecki assaulted me. He pinned me against a bar, he put his hand up my skirt and his fingers in my underwear. He grabbed and twisted. He leaned in and insisted that I sleep with him. “It could be innocent,” he said.
Months of struggle, self doubt and multiple rejections later: received an internship offer from
@google
for the student researcher role in machine translation!
Things do work out, even though i was ready for a less glorious simple one. So excited and grateful! 🤍
Honestly, rejections are so difficult. Every rejection you tell yourself that something better awaits you, until something you thought was better also rejects you. Whatta weird loop. I will still think everything worked out just fine but while it doesn’t, it’s difficult.
The last two years, of being at Stanford and Google, made me believe academically women were as trusted in their ability to do ML/research as their counterparts, and maybe I was just in the wrong bubble before. The past couple of weeks has taught me I was wrong.
Hmph it's happening!
Masters in Symbolic Systems, let's do this! Excited to study and research in CS, NLP, Psych and philosophy of mind ✨
If you're at Stanford or nearby or a recruiter, would love to connect!
I’m tired of the toxic disproportion of men in STEM, particularly computer science, unfortunately my major. This makes it easier for any toxic tech bro to treat me as invisible if I don’t submit to them. The other relatively non-toxic ones extend their silent support in isolation
Every time i explain my work at Google or my research at Stanford to someone, I’m inundated with so much happiness. I hope to always be this excited about my projects (even though they might not go as great in some temporary phases).
Being an international student can be so difficult sometimes. First Diwali, now the cousin I'm closest to is getting married, while I struggle between food, courses and planning the next quarter 😭
Sometimes all I wanna do is rush back home and attend the smaller moments.
Honestly, rejections are so difficult. Every rejection you tell yourself that something better awaits you, until something you thought was better also rejects you. Whatta weird loop. I will still think everything worked out just fine but while it doesn’t, it’s difficult.
Feels so nice to meet with professors who you vibe with on not just research ideas and ideologies but who are also transparent and truly care about your progress, time and energy. 😭✨🧿
As I looked at all my publications yesterday, I realised how far I've come. From a largely spaced badly formatted first paper to a sophisticated looking deeply researched last paper - genuinely felt very nice. Would highly recommend doing so!
In my undergrad, from the limited exposure around me, I did not feel I had a lot of interesting choices so it was easy to pick a fav & stick to it.
Coming to Stanford I was drowned in decision dilemma since so many seemed so interesting. Naturally, I developed commitment issues.
lol kinda tired of all these AI startups using cognitive in their names/websites, with a central image of a brain in website header only to find out that they do generic machine learning and software engineering for industry and nothing neuro/brain/cogsci related.
To all the undergrads in India: a gem of a person and researcher, Danish, is relocating back to India. Reach out to him if you’re interested in applicative NLP. Anyone would be lucky to have him as their PI/collaborator 🤌🏻
I am beyond thrilled to share that I'll be starting as an assistant professor at the Indian Institute of Science (IISc), Bangalore in April 2023.
I couldn’t have been luckier—I'm grateful for the support of many kind mentors, peers, students, friends and family members. (1/4)
Felt prey to the typical grad student experience where you overload yourself with work until you cannot breathe. Took a 1 min reflection break to think why I’m doing the research I’m doing; 5 hours since, never been so happily productive and need I say dancy:)
Highly recommend!
Today I am so glad I tweeted this. A girl reached out, and we were (hopefully) able to prevent a potential repetition of something similar by the same professor. Twitter is a bful space 🌼
In my junior year of undergrad, I worked with a reputed professor from a very esteemed university of India. There would be talks about my alcohol preferences in the 8 months of internship, but I took it in light humour. After my project, the prof wanted me to work on a project..
While I really appreciate the support, I would sincerely ask people to let me deal with this at my pace. I would take actions if and when I feel comfortable and would appreciate people not jumping ships and acting on my behalf. Thank you!
And the two month break ends now. It was difficult to not do anything and just relax, sleep, play and spend time with my family but I think it was also required. Let's get back to work now 💪🏻
Today’s learning: if you want something, just ask.
Feels like a weird motivational cliché quote, but as a person who is afraid of rejections and being an inconvenience, it feels good to take a step (:
Should I create a anonymous account to fangirl over chance encounters with professors?
I was converting an R code to Python and i saw Chris Manning pass by and turn his head twice to me AND I CANT…
how do you breathe?
Just wanna go back to my shell and scream. 😭😭
Today might be my (second) happiest day. I've been feeling lost with where my problem statement lies and how to go about it. And today I got to talk to Dan (Jurafsky) about it. I feel so lucky and overwhelmed but also great.
Twitter has been my safe space for anything research..
The past two years have been keeping it limited to my close friends for the "lack of evidence", or the fear of him sabotaging my grad school applications (he was well connected) or random excuses I made. But this thread reminded me too much and too strongly to stay silent.
Ok enough. Been ignoring hundreds of messages, comments and retweets about this. How difficult is it to accept a good gesture without any ulterior motives or hidden emotions and just out of plain bful friendship? Stop ruining one of the last male fships I'm left with ugh
Idk if it is the right time or place to put it out there but it has eaten a part of me, made me scared of a lot of things, and I just wanted to put it out there and hopefully move on as I begin the next chapter of my life.
As for what’s next, I’m super stoked to design the future of LLMs with
@realSharonZhou
,
@GregoryDiamos
and the wonderful people
@LaminiAI
🦙🎉
A month in, I am amazed by the warmth and knowledge of everyone around me! 💛
A course on what we are building:
New short course on Fine-tuning LLMs! Many developers are moving beyond only prompting, to also fine-tuning LLMs - that is, taking a pre-trained model and training it further on your own data, which can deliver superior results inexpensively. In this course,
@realSharonZhou
, CEO
Now I am beginning to stick to one while gaining knowledge about others. How? Work in the main interest, take coursework in others, one at a time and resume community work. Seems very simple but took me a while to figure out. Things are beginning to feel better now:)
As the application season brews, I receive a couple of DMs each day seeking help and advice. As much as I desperately want to help you folks, quarter system makes it so difficult to do so. So I'm planning to organise a session in early December to help answer some questions..
Finally done w
#EACL2021
, it was one of the most enriching and amazing weeks. I met some amazing researchers, talked to professors I only saw on twitter, presented my first poster at a major conf and had some very interesting conversations around interpretable AI.
Social media has subtly forced me to look back on 2021. This year gave me my top two moments of life, yet. Getting into my dream program at an unbelievably dream university and however lifeless that sounds, meeting a professor I’ve only insanely fangirled over.
We have reached an agreement in principle for Sam Altman to return to OpenAI as CEO with a new initial board of Bret Taylor (Chair), Larry Summers, and Adam D'Angelo.
We are collaborating to figure out the details. Thank you so much for your patience through this.
A very interesting panel at the workshop of
#FoundationModels
today. great points on the importance of data, architecture revamp, evaluation metrics and explainability, and grounding. Thank you for the insights
@YejinChoinka
@prfsanjeevarora
@natschluter
and other professors.
I’m tired. It feels easier to swing to a field with more representation than having to explain myself to everyone all the time and still feeling unsafe, unheard and spoken over.
Humans of AI is one of the most enticing series ever. I've been promising myself to watch it for the longest time now but maybe seeing Yejin on it was the final motivation I needed!!
S2E20 is out!
Yejin Choi
@YejinChoinka
(
@uwcse
,
@allen_ai
) on Humans of AI: Stories, Not Stats.
Yejin talks about where she comes from, living life like in a game environment, thinking in vector spaces, finding her true self, and lots more.
[1/n]
While describing an action (like a project, a course, or a direction) in your Statement of Purpose, there are so many realities to describe the intuition behind the move. What differentiates a good SOP imo is how coherent one's chosen reality is with their overall goal.
Idk what’s the moral of this thread but it feels too suffocating to continue suffering silently while the world continues breezily.
And then people question why men’s identity is about their work and women’s about being a women; why the bar “has to be lowered” for women; etc.
I joined Twitter because I wanted to be aware of the academic opinions from top scientists across the globe and broaden my purview. Now I see humdrum and mostly ungeneralizable arguments about something as big as choosing a PhD, and I'm afraid for the kids reading them.
Catching up on personal emails and found this! How'd I get so lucky 😭
I miss teaching, but also I had the best students 💕
For context, Symbolic Systems was my Masters program at Stanford and this course was symsys1: Minds and Machines where we discuss CS x Lang x Psych x Phil
Just a little weird to me that the internet decided who the “good guys” were based on the narrative only one side expressed. We may never know the other side, for whatever reasons legal or otherwise, but just weird to me
How are students (and not professors) getting 3 or 4 paper accepted at ICRA or ICCV or any similar conference? How do you produce the quality with so much quantity. My guess is we can't.
This numbers game tbh makes me feel lost about where we are heading to.
A friend accidentally closed 500+ tabs in 10+ windows (in a windows laptop on chrome). Any idea how to bring it back? (ctrl+shift+T isn't working and searching through history is impossible) 🥲
(is there a community on twitter I can tag which can help?)
Um, I love ACMI's research but smth about if I "suck energy", the fear that I didn't "do anything", or the expectation of shooting to moon and forcefully arguing for what you think is right makes me feel..unsafe?? Idk
Research experience is great, published papers can be impressive, but (generally) they are neither necessary nor sufficient for joining our lab. Some of our criteria:
Such an important thread.
I remember when I attended
@mcxfrank
's lab meeting, it was the first time I saw people discuss achievements in their personal life, including getting a cat. It felt like people in the meeting were human and a community, smth that was rare in CS academia.
I have seen so many threads recently bemoaning the working culture in academia, and I fully agree with many things. BUT: y'all know you are academia and it is on you to come together and fix it, right?
I have a radical and simple first proposal: Don't work on the weekend. (1/9)
So now when they focus on working on their startups and courses, I am still processing and dealing with the fresh invisibility and thinking if i am the problem and did smth wrong:)
For "female hygiene" for which he asked I click pictures of my vagina before and after menstruation and plot histograms of it (??). Since I wasn't comfortable and honestly conflicted about the thin line b/w research and woke-ness, I delayed. This was followed by him constantly..
I made a random excuse and hung up. But I froze, not sure what to do. The next day I saw him being hailed by the organisers and attendees.
Somehow I made it back to my home but I could not look at him throughout.
This quarter I am leading sections where I get to teach SymSys 1 (Mind and Machines) to small groups of students. I had almost forgotten how hard but enjoyable teaching was. Yday was the first lecture and I rolled out feedback forms; one student literally wrote this 😭
So while all of them bond stronger, my past trauma is corroded further and it is brushed off as an overthinking overreaction. So everyone comes back from spring break as happy, relaxed, closer to one another and I with regular breakdowns.
Sometimes I miss brain spilling in the empty abyss my twitter was
Like how an amazing professor whose work I very much adore visited the Stanford campus this week and I saw the email the night he visited. I let out a (small) scream at 11 pm
As the grad school applications gear up, a common question is how do you choose schools. I wanted to offer my 2 cents on it.
For a person who is indecisive by nature, I feel it has been fairly easy to make professional decisions for me. How? People.
I love February because it has exactly 4 weeks - so satisfying.
Proposal: Put all extra days from the remaining months into a new month and the lingering extra day is the new year day. For leap year we get another new year day to introspect on the past 4 years.
Being from an underrepresented community, I did not have many people to seek help from when I was applying last year. Then I saw this and I can say with confidence that GASP was the best thing that happened to me last year. So grateful to
@shrimai_
and co for this. Apply!
Historically, people from underrepresented backgrounds have lacked access to guidance from people familiar with the higher degree application process. GASP matches applicants to LTI programs with grad students who provide feedback on their applications
I’m going to be in Seattle tomorrow onwards! If you’re there too (
#NAACL2022
) let’s catchup (anything from machine learning to cognitive neuroscience, life and art) during, or after the hours! 💃🏻
Pls also send interesting talks info (feel free to plug your work!). Very hype!
An interesting argument by
@ev_fedorenko
in her
#EMNLP2021
keynote: humans use language for communication, not for thought—and so language models should not be judged for their ability to reason.
I would add: we should also not use lang.models as proxies for cognitive abilities.
So I had 3 glasses of champagne and got to Cal Anderson with a cake my mom bought me and an also drunk couple asked about my thesis and the dude was a constructivist I swear to god. So I fell into the grass and we spoke for like 20 minutes and there was a dog 🐕
Yes, the year was difficult because drastic changes brought back a lot of insecurities but i have hope from 2022 to be a year of personal and professional growth.
Until then, grateful to the friends and family i have and for this life and experiences. ❤️
I've used twitter as a space to rant, fangirl and share cool papers. This was one such moment. I didn't expect it to blow up (stupid assumption in hindsight). But welcome everyone to my thought express, feels weird to have so many of you looking at me but guess that's how it is
Asking if I clicked "pictures of the event". I refused, following which he didn't look at my paper for an upcoming conference. The conference was 1500kms away from my home and he was the only person I knew there.
For the past few days, I have been asked multiple times about the 'Predoctoral Researcher' role and its hiring process at
#Google
#Research
, India. So, I have compiled the answers and my experience in this blog post: . Hope this helps the upcoming batches.
This summer was so enriching, on so many levels, personally and professionally, I’m excited to get back into the business and write a blog about it. (Also about my selection process - on popular request). Stay tuneeed (feeling like an influenza while I write this but ya)
Can not detest and downvote this chain of thought enough. If you’re a budding women in stem, please disregard this advice and utilize every little opportunity you get (and if you’re an organizer, please don’t stop creating such opportunities) ty
It is quite hard to criticize any diversity event because many times the organizers take it personally and there remains no chance of having a healthy debate. So let me tell you objectively why I think diversity events are causing more harm than good, in the long term.
1/
Announcing C4AI Command R+ open weights, a state-of-the-art 104B LLM with RAG, tooling and multilingual in 10 languages.
This release builds on our 35B and is a part of our commitment to make AI breakthroughs accessible to the research community. 🎉
After my paper presentation I asked him where we could have dinner (since the hosted dinner tasted very bad and he offered to get dinner outside if I didn't like the one the hosts served.)
Thank you for the most beautiful words on this, in replies and messages. Internet is sometimes an unexpectedly kind world ❤️
Took a two day break and now back to work, things will get better 💪🏼
Honestly, rejections are so difficult. Every rejection you tell yourself that something better awaits you, until something you thought was better also rejects you. Whatta weird loop. I will still think everything worked out just fine but while it doesn’t, it’s difficult.
We give such great importance to being strongly opinionated that we forget to appreciate the art of hearing other people and the people who exercise it.