Adriano D'Alessandro
@adrian-dalessandro.bsky.social
| Computer vision researcher
| Computer science PhD candidate @ SFU
| More: https://dalessandro.dev/
I like to count things and periodically I work on applications in plant agriculture + ecology.
Follow for stale political hot takes.
Free Palestine 🇵🇸
| Computer science PhD candidate @ SFU
| More: https://dalessandro.dev/
I like to count things and periodically I work on applications in plant agriculture + ecology.
Follow for stale political hot takes.
Free Palestine 🇵🇸
Pinned
LookAlikes Dataset
LookAlikes is a fine-grained object-counting dataset for evaluating the generalization of class-agnostic counting methods. Each image contains visually similar objects, from which only a subcategory m...
dalessandro.dev
Want to evaluate open-world object counting in a fine-grained setting? We’re excited to release the LookAlikes dataset, which is a test-set-only benchmark where images contain objects from multiple visually or semantically similar categories! #ComputerVision #DeepLearning #AI #Counting #OpenDataset
AGI will be achieved when an AI can successfully take a duvet cover out of the laundry and put the cover on a blanket. That's it.
November 2, 2025 at 4:08 AM
AGI will be achieved when an AI can successfully take a duvet cover out of the laundry and put the cover on a blanket. That's it.
One idea for improving reviews at conferences is to assign one reviewer as a champion by default. Their role is to find every possible reason to accept the paper.
I find reviewers often spend most of their time focusing in on what is bad in a paper. This is useful! But more often than not
I find reviewers often spend most of their time focusing in on what is bad in a paper. This is useful! But more often than not
October 22, 2025 at 6:21 AM
One idea for improving reviews at conferences is to assign one reviewer as a champion by default. Their role is to find every possible reason to accept the paper.
I find reviewers often spend most of their time focusing in on what is bad in a paper. This is useful! But more often than not
I find reviewers often spend most of their time focusing in on what is bad in a paper. This is useful! But more often than not
Now that the #AAAI2026 discussion is mostly complete, I want to give my final assessment of the AI summary from the perspective of a reviewer:
1. The positives: It functioned like a literature review + found minor errors and inconsistencies in the paper. Really useful as a high-level sanity check.
1. The positives: It functioned like a literature review + found minor errors and inconsistencies in the paper. Really useful as a high-level sanity check.
October 20, 2025 at 8:40 PM
Now that the #AAAI2026 discussion is mostly complete, I want to give my final assessment of the AI summary from the perspective of a reviewer:
1. The positives: It functioned like a literature review + found minor errors and inconsistencies in the paper. Really useful as a high-level sanity check.
1. The positives: It functioned like a literature review + found minor errors and inconsistencies in the paper. Really useful as a high-level sanity check.
This is an VERY important point, and I wonder how much knowledge we have lost because an interesting research path is cutoff due to early results not outperforming big models on some benchmarks. We are "expecting skyscrapers to learn to fly if we build them tall enough", while punishing other ideas.
So all these big models - commercial models I must add - are really nice tools, impressive engineering, but the additional knowledge that they bring to the community is minimal in my opinion.
There is no contradiction in being in awe before Sora2's results and thinking that...
There is no contradiction in being in awe before Sora2's results and thinking that...
October 3, 2025 at 4:52 PM
This is an VERY important point, and I wonder how much knowledge we have lost because an interesting research path is cutoff due to early results not outperforming big models on some benchmarks. We are "expecting skyscrapers to learn to fly if we build them tall enough", while punishing other ideas.
Want to evaluate open-world object counting in a fine-grained setting? We’re excited to release the LookAlikes dataset, which is a test-set-only benchmark where images contain objects from multiple visually or semantically similar categories! #ComputerVision #DeepLearning #AI #Counting #OpenDataset
LookAlikes Dataset
LookAlikes is a fine-grained object-counting dataset for evaluating the generalization of class-agnostic counting methods. Each image contains visually similar objects, from which only a subcategory m...
dalessandro.dev
September 11, 2025 at 8:52 PM
Want to evaluate open-world object counting in a fine-grained setting? We’re excited to release the LookAlikes dataset, which is a test-set-only benchmark where images contain objects from multiple visually or semantically similar categories! #ComputerVision #DeepLearning #AI #Counting #OpenDataset
Why are there only 2 major CV conferences a year? If I want to submit a deep learning paper, there is NeurIPS/ICLR/ICML/AAAI. For a CV paper, I only have CVPR and ECCV|ICCV (Nov/Mar deadlines). This year there were 10K CV papers submitted to AAAI and 11K to ICCV. There should be a July CV deadline
August 28, 2025 at 6:21 PM
Why are there only 2 major CV conferences a year? If I want to submit a deep learning paper, there is NeurIPS/ICLR/ICML/AAAI. For a CV paper, I only have CVPR and ECCV|ICCV (Nov/Mar deadlines). This year there were 10K CV papers submitted to AAAI and 11K to ICCV. There should be a July CV deadline
GPT5 still can't solve this visual puzzle, which is interesting!
August 20, 2025 at 6:32 AM
GPT5 still can't solve this visual puzzle, which is interesting!
I guess we know what's left for GPT6
August 8, 2025 at 1:13 AM
I guess we know what's left for GPT6
I've been reading a bit of Byung-Chul Han, recently, and I'm starting to think the male loneliness epidemic is just a byproduct of digital ecosystems destroying our ability to form coherent narratives about our lives. Dating is reduced to listing of algorithm friendly facts. We convert ourselves to
August 2, 2025 at 6:48 AM
I've been reading a bit of Byung-Chul Han, recently, and I'm starting to think the male loneliness epidemic is just a byproduct of digital ecosystems destroying our ability to form coherent narratives about our lives. Dating is reduced to listing of algorithm friendly facts. We convert ourselves to
Paper idea: Adversarial images but for people. When shown this image of a real sand sculpture, organic intelligence tends to identify it as fake.
August 2, 2025 at 6:27 AM
Paper idea: Adversarial images but for people. When shown this image of a real sand sculpture, organic intelligence tends to identify it as fake.
On deep learning and late capitalism: I was listening to a talk by Mark Fisher on the "Slow Cancellation of the Future" and he makes this point that the internet has made the past available in a decontextualized way where it becomes an aesthetic choice rather than anything coherent and organized.
June 29, 2025 at 7:15 PM
On deep learning and late capitalism: I was listening to a talk by Mark Fisher on the "Slow Cancellation of the Future" and he makes this point that the internet has made the past available in a decontextualized way where it becomes an aesthetic choice rather than anything coherent and organized.
Damn 😥
Elevated heart rate? Focus in shambles? Ah, yes, it's decision day for #ICCV2025 🤢🥶
June 25, 2025 at 6:51 PM
Damn 😥
The predictive power of the 20 year nostalgia cycle is unparalleled.
June 22, 2025 at 4:13 PM
The predictive power of the 20 year nostalgia cycle is unparalleled.
o4-mini-high doesn't know about "Loss"
June 3, 2025 at 9:09 PM
o4-mini-high doesn't know about "Loss"
This is an interesting visual puzzle, which o4-mini-high fails to solve. It never considers the orientation of the letters with respect to each other and thus that there must be one letter that exists twice on the cube.
May 29, 2025 at 9:36 PM
This is an interesting visual puzzle, which o4-mini-high fails to solve. It never considers the orientation of the letters with respect to each other and thus that there must be one letter that exists twice on the cube.
Did gpt-4o get a bunk update recently? I gave it this riddle and it fell through the floor.
May 25, 2025 at 6:25 AM
Did gpt-4o get a bunk update recently? I gave it this riddle and it fell through the floor.
Now you're thinking strategically and hitting on one of the core contentions of the review process! Let's dig into it --
Today we announced a pilot program that strategically
incorporates Large Language Models to enhance the academic paper review process for the AAAI-26 conference. Read more:
aaai.org/aaai-launche...
incorporates Large Language Models to enhance the academic paper review process for the AAAI-26 conference. Read more:
aaai.org/aaai-launche...
AAAI Launches AI-Powered Peer Review Assessment System - AAAI
AAAI today announced a pilot program that strategically incorporates Large Language Models (LLMs) to enhance the academic paper review process for the AAAI-26 conference.
aaai.org
May 17, 2025 at 9:06 PM
Now you're thinking strategically and hitting on one of the core contentions of the review process! Let's dig into it --
The USA is engaging in religiously-motivated organ harvesting targeting woman. Like, just out in the open.
The state of Georgia is using the body of a brain-dead woman as an incubator. Because she was 9 weeks pregnant when she died, and their abortion ban is from about 6 weeks, they are sustaining her on life support, without input from her family, until 32 weeks of gestation. A true dystopian nightmare.
May 16, 2025 at 9:08 PM
The USA is engaging in religiously-motivated organ harvesting targeting woman. Like, just out in the open.
Asking ChatGPT to convert the children's drawing (left) to an anime character (middle) vs. a human attempt (right).
May 6, 2025 at 3:04 AM
Asking ChatGPT to convert the children's drawing (left) to an anime character (middle) vs. a human attempt (right).
ChatGPT has gotten substantially worse as a critic. It really wants to tell you how good and smart you are. Excuse me but work is NOT good and I'm not smart enough to notice why. Please do your JOB. Go to the ICLR openreview and find the lowest rated paper you can. Ask ChatGPT is it would accept the
April 29, 2025 at 9:06 PM
ChatGPT has gotten substantially worse as a critic. It really wants to tell you how good and smart you are. Excuse me but work is NOT good and I'm not smart enough to notice why. Please do your JOB. Go to the ICLR openreview and find the lowest rated paper you can. Ask ChatGPT is it would accept the
There’s a pattern in deep learning I think of as “first to a dead-end.” Defining a novel problem often gets you a flashy paper, but follow-ups become hard to justify without being labeled incremental. So the problem is quietly abandoned. It’s a product of an oversaturated field, where it's safer to
April 16, 2025 at 2:41 AM
There’s a pattern in deep learning I think of as “first to a dead-end.” Defining a novel problem often gets you a flashy paper, but follow-ups become hard to justify without being labeled incremental. So the problem is quietly abandoned. It’s a product of an oversaturated field, where it's safer to
I waited 2 hours to vote in the Vancouver byelection yesterday. I'm absolutely ecstatic to see a landslide for socialist and progressive candidates Sean Orr and Lucy Maloney.
April 6, 2025 at 2:54 PM
I waited 2 hours to vote in the Vancouver byelection yesterday. I'm absolutely ecstatic to see a landslide for socialist and progressive candidates Sean Orr and Lucy Maloney.
Why are the ICCV submissions in my review pile so good 😭 you all got me sweating for my submission.
March 28, 2025 at 6:34 PM
Why are the ICCV submissions in my review pile so good 😭 you all got me sweating for my submission.