Ashish

Ph.D. student at Computer Vision Lab, UMass Amherst

[Ashish @ashish9494]

@SamuelAlbanie was the average human reading a book on phone? does that make a difference? (genuinely asking for myself as i get lot of flack for reading books on phone🙈)

[Chris Albon @chrisalbon]

From my private notes on 2023-12-31:

Looking back from around 2018 until now, the biggest lesson I will take into 2024 is this:

Focus my attention on things I want to accomplish AND I have significant control over.

People have said roughly the same thing in many ways, such as "Play your own game.", but it all boils down to focusing on the things that matter to you and that you can affect in a meaningful way.

People always want your attention. This has been true forever, it has become more true with social media, and it will be even more true with the US presidential elections in 2024. Sometimes people want your attention to monetize it (ads, etc.), others want your attention to try to influence public perception around some event (elections, etc.), and often it is a bit of both.

In 2020, when the pandemic hit, I felt this need to follow these topics in excruciating detail. What is the COVID rate in SF vs. the US today? Is the trend line going down? What is the latest rumor about the vaccine?

At the time, I left like this was a good thing -- I was being an "informed citizen" and I was bearing some emotional witness to whatever was happening.

But I was wrong. It wasn’t me being an informed citizen. It was me toxically ruminating over things that I couldn’t control. There was nothing I could write, or code that I could deploy that would help in any meaningful way, and so I just sat there obsessively reading about it day after day.

The hundreds of hours I spent ruminating about COVID did not make the vaccine come one millisecond faster or make the pandemic wind down faster. The situation was not something I had any significant control over. Other people had significant control, but not me.

And the truth is that my ruminating had real costs. Sitting for over a year being stressed about something but also lacking control over the thing is a recipe for a very bad mental state. I drank more, I ate worse, I gained a lot of weight, I stopped exercising, I became depressed, all of which became their own problems. All because I felt it was important or in some way helpful to obsess about something I had effectively zero control over.

Coming out of that, the lesson I took away is that there are things I have significant control over in my life, because of my skill set, location, experience, and relationships. And then there is a 100,000x larger set of things that are just too big or outside my area of influence for me to have any significant impact on.

So my lesson for 2024 is to let the things I don't have meaningful control over to pass right through me. Don't ignore them, don't dig my head in the sand, but acknowledge them and then purposely avoid directing the Eye of Sauron that is my obsessive nature on them.

Instead, focus my energy on things inside my area of significant control. What are the things I want to see that I can have a significant control over? What can I personally do to advance those things in a meaningful way? What code can I release open source? What project can I start? What two people can I introduce?

[Rohan Pandey @rohan99pandey]

I worked with a student on a paper. They submitted a small extension, copying multiple sections, re-using codes, and pretty much the entire idea - and submitted (& got accepted to ACL SRW) as a sole author. 4 people worked on this, is there something I can do about this?

[Ashish @ashish9494]

@ylecun @elonmusk @erikbryn this has exactly been my argument when someone disses Biden administration over economy!

[Ashish @ashish9494]

@CVPR @jcniebles it takes me 12 seconds to just comprehend the
title itself🙃

[Ashish @ashish9494]

@y_m_asano thank god i am not the only one facing this problem!

[Ashish @ashish9494]

@arankomatsuzaki this made me really happy!

[Ashish @ashish9494]

@ftm_guney that is true😁. By the way, if your students are interested, another interesting case where this model suffers sometimes is very very long term occlusion. For example a pedestrian on a sidewalk and getting occluded by bus for a long time.

[Ashish @ashish9494]

@ylecun @DrJimFan @RichardSSutton what is your opinion on designing better objective functions for training models?

[Grant Van Horn @Grant_Van_Horn]

[1/4] 📢 PhD positions open for Fall '24 in my new UMass Amherst lab! Apply by Dec 15. Details: https://t.co/qcfaDioqGl

[Ashish @ashish9494]

@oriane_simeoni that makes sense. thank you for the response😁

[Ashish @ashish9494]

@oriane_simeoni should papers on open world object proposals be included here? or is it off-topic?
For example this one?

https://t.co/in8szG0qMG

[Ashish @ashish9494]

@pcascanteb congratulations!!😁😁

[F. Güney @ftm_guney]

Brad Pitt on awards and stuff: “I’ve been banging away at this thing for 30 years. I think the simple math is, some projects work and some don’t. There’s no reason to belabor either one. Just get on to the next.”

[Kwang Moo Yi @kwangmoo_yi]

A few things I ask my students to do when writing reviews:
1. Remove all adverbs
2. Try removing all adjectives
3. You are not the judge -- you are either the prosecutor or the lawyer.
4. Be evidence-based.
5. Be kind -- there's someone's <1y of life behind it.

[Ashish @ashish9494]

@gowthami_s same!😅😂

[Ashish @ashish9494]

@ZezhouCheng @MajiSubhransu @georgiagkioxari @jampani_varun @HuaizuJiang @gadelha_m yaaay! congrats zezhou!! 🥳

[Shashank Tripathi @sha2nk_t]

<-- Follower counts -->
GoogleAI - 1.9M
OpenAI - 2.3M
MetaAI - 409K
Famous CV Researchers - ~15K
Me - 448
Let the competition begin. 🥲

[Ashish @ashish9494]

@cristiancanton where can i get the full video to try out things? is it on youtube??