Jason Corso

Quit by Annie Duke ...a real life decision making read.

Awesome! What type of work is he looking for?

Sure, here I am thinking, yes, but, umm, where did June 2020 go.... :)

The world according to Elon. 🤦

What was final attendance number at #CVPR2025? @cvprconference.bsky.social

Auto-Labeling Data for Object Detection Great labels make great models. However, traditional labeling approaches for tasks like object detection have substantial costs at scale. Furthermore, alternatives to fully-supervised object detection...

- Understanding the tradeoffs of confidence thresholds enables better tuning of auto-labeling pipelines, prioritizing overall model effectiveness over superficial label cleanliness.

Read the full paper from the @Voxel51 team: arxiv.org/abs/2506.02359

💡Why this matters:
- Annotating large visual datasets can be done 100,000x cheaper and 5,000x faster with public, off-the-shelf models while maintaining quality.
- Highly accurate models can be trained at a fraction of the time and cost of those trained from human labels.

High-confidence labels (0.8–0.9), while appearing cleaner, consistently harmed downstream performance due to reduced recall.

Understanding this balance enables better tuning of auto-labeling pipelines, prioritizing overall model effectiveness over superficial label cleanliness.

📊 Higher confidence threshold isn’t always better.

This is an interesting one. Somewhat counterintuitively, setting a relatively low confidence level (ɑ ≈ 0.2) for auto labels maximized the precision and recall of downstream models trained from auto-labeled data.

📊For less-represented classes, models trained from auto labels sometimes performed even better.

The image below compares a human-labeled image (left) with an auto-labeled one (right). Humans are clearly, umm, lazy here :)

On COCO, auto-labeled models reached mAP50 of 0.538 compared to 0.588 for human-labeled counterparts, demonstrating competitive real-world performance.

📊 Comparable accuracy to human labels.

The mean average precision (mAP) of inference models trained from auto labels approached those trained from human labels.

On VOC, auto-labeled models achieved mAP50 scores of 0.768, closely matching the 0.817 achieved with human-labeled data.

The findings:
📊 Massive cost and time savings.

Using Verified Auto Labeling costs $1.18 and 1 hour in @NVIDIA L40S GPU time, vs. over $124,092 and 6,703 hours for human annotation.

Read our blog to dive deeper: link.voxel51.com/verified-auto-labeling-tw/

- Trained downstream inference models on both the auto-labeled and human-labeled data and compared their performances
- Translated our research into a new annotation tool called Verified Auto Labeling

🛠 What we did:
- Used off-the-shelf foundation models to label several benchmark datasets
- Evaluated these labels relative to the human-annotated ground truth

Annotation is frequently a critical bottleneck in computer vision, often driving data labeling costs into the millions.

Our goal with this experiment was to quantitatively evaluate how well zero-shot approaches perform and identify the parameters and configurations that unlock optimal results.

Exciting CVML result from the @voxel51.bsky.social ML team: zero-shot object detectors are 100,000x cheaper yet rival humans

The zeitgeist claims zero-shot labeling is here but no one measured it. We did. 95%+ performance of human labels 100,000x cheaper & 5,000x faster

arxiv.org/abs/2506.02359

have there been instances of CV/ML conferences where the pre-rebuttal review did NOT have the recommendations published to the authors? I feel like the rebuttal process may be more fruiftul if there was more focus around the content only than the content biased by scores.

Trump Administration Live Updates: President Demands Walmart Keep Prices Low Despite Tariff Costs

In the latest from idiot-land...how one can be so disconnected from reality is beyond thought. Peoples' livelihoods and, umm, peoples' kitchen tables are in play. Trump Administration Live Updates: President Demands Walmart Keep Prices Low Despite Tariff Costs www.nytimes.com/live/2025/05...

The nightmare continues. And AAAI of all communities.

What's happening is awful. Is anyone aware of a law firm that has stepped up to cohesively collate responses against these illegal actions? Individual universities will likely handle things wildly differently, and, oh, have you heard? they're distracted now with other illegalities

be ready to debug debug debug.

Boom. Indeed.

It's the last week of @iccv.bsky.social reviewing, a bit before @neuripsconf.bsky.social reviews, among others.

Check your mindset as you review. Are you biased to
(1) finding reasons to reject the paper?
(2) finding reasons to accept the paper?
(3) finding reasons to understand the paper?

JPMorgan's Dimon says economy is facing 'considerable turbulence' JPMorgan Chase CEO Jamie Dimon said Friday that the economy faces "considerable turbulence" as his bank braces for future loan losses.

It definitely takes a big bank CEO saying it for it to be true. GMAB finance.yahoo.com/news/jpmorga...

Congratulations @kostaspenn.bsky.social!

same. Found it enjoyable, but definitely missed many references. Glad there was some notion of a plot!

I'm really curious what that "community" impact is going to be if this happens.

According to the FBI (what's left of it), this is the first known case of an AI agent "extrapolating" its way into full-blown identity theft. 🧐🧐🧐

The culprit? A rogue AI agent with a flawed sense of "saving money." Its system prompt told it to save money for *customers*, but somehow, it decided the best way was to steal my identity to do so. 🤖 What guile!