Erhan Bas

58 posts

Erhan Bas

Erhan Bas

@erhnbs

ML, Computer vision, machine learning, neuroscience, physics and games.

Katılım Haziran 2013
197 Takip Edilen59 Takipçiler
Erhan Bas retweetledi
AK
AK@_akhaliq·
Scale AI, NYU, and UC Berkeley present Let's Go Shopping (LGS) Web-Scale Image-Text Dataset for Visual Concept Understanding paper page: huggingface.co/papers/2401.04… Vision and vision-language applications of neural networks, such as image classification and captioning, rely on large-scale annotated datasets that require non-trivial data-collecting processes. This time-consuming endeavor hinders the emergence of large-scale datasets, limiting researchers and practitioners to a small number of choices. Therefore, we seek more efficient ways to collect and annotate images. Previous initiatives have gathered captions from HTML alt-texts and crawled social media postings, but these data sources suffer from noise, sparsity, or subjectivity. For this reason, we turn to commercial shopping websites whose data meet three criteria: cleanliness, informativeness, and fluency. We introduce the Let's Go Shopping (LGS) dataset, a large-scale public dataset with 15 million image-caption pairs from publicly available e-commerce websites. When compared with existing general-domain datasets, the LGS images focus on the foreground object and have less complex backgrounds. Our experiments on LGS show that the classifiers trained on existing benchmark datasets do not readily generalize to e-commerce data, while specific self-supervised visual feature extractors can better generalize. Furthermore, LGS's high-quality e-commerce-focused images and bimodal nature make it advantageous for vision-language bi-modal tasks: LGS enables image-captioning models to generate richer captions and helps text-to-image generation models achieve e-commerce style transfer.
AK tweet media
English
0
26
138
22.1K
Erhan Bas retweetledi
Scale AI
Scale AI@scale_AI·
Scale’s paper “Detecting and Preventing Hallucinations in Large Vision Language Models” has been accepted for publication in AAAI 2024! 🧵 arxiv.org/abs/2308.06394
English
1
9
29
3.8K
Hongkui Zeng
Hongkui Zeng@HongkuiZeng·
Deeply humbled by this tremendous honor! I am incredibly grateful for the extraordinary opportunity of working at the @AllenInstitute alongside so many amazing colleagues.
allen institute@AllenInstitute

Our own @HongkuiZeng has been elected to @theNAMedicine! One of the highest honors in the health and medicine field, the election recognizes Zeng for her outstanding professional achievements and commitment to service. alleninstitute.org/news/hongkui-z…

English
23
10
227
51.6K
Erhan Bas retweetledi
Adam Selipsky
Adam Selipsky@aselipsky·
This innovative area of deep learning research we’re calling “Graceful AI” understands there are multiple goals in a problem and explores how to make newer AI models compatible with previous systems. Check out a great deep dive into the project: zdnet.com/article/amazon…
English
5
19
62
0
Erhan Bas retweetledi
Nature Methods
Nature Methods@naturemethods·
SNT: Following in the footsteps of Single Neurite Tracer, the SNT toolbox is useful for neuromorphometry and connectomics and provides various analysis, visualization, quantification and modeling tools. @tiago_tferr nature.com/articles/s4159…
Nature Methods tweet media
English
1
9
29
0
Erhan Bas retweetledi
Erhan Bas retweetledi
Eric Topol
Eric Topol@EricTopol·
Largest #LongCovid cohort @TheLancet just published: 6 months post-hospital, —76% still had ≥ 1 symptom —63% fatigue/muscle weakness —sleep, anxiety, depression common —Sig % w/CT lung abnormalities —13% had abnormal kidney function thelancet.com/journals/lance…
Eric Topol tweet mediaEric Topol tweet media
English
92
2K
3.3K
0
Erhan Bas
Erhan Bas@erhnbs·
@perlman Wow Eric... So sorry to hear that. Hope that you recover fast.
English
0
0
1
0
Erhan Bas retweetledi
Peyman Milanfar
Peyman Milanfar@docmilanfar·
(1/5) One of the most surprising and little-known results in classical statistics is the relationship between the mean, median, and standard deviation. If the distribution has finite variance, then the distance between the median and the mean is bounded by one standard deviation.
Peyman Milanfar tweet media
English
11
633
2.7K
0
Erhan Bas retweetledi
Johan Winnubst
Johan Winnubst@JohanWinn·
The answer to: "Why don't you just use EM to reconstruct neurons in the mouse brain like they do in in the fly?"
English
4
163
617
0
Erhan Bas retweetledi
Dagmar Kainmueller
Dagmar Kainmueller@kainmuellerd·
We’re looking for a PhD student to work on Deep Learning from Sparse Annotations for Cell Segmentation in the HEIBRiDS grad school in Berlin (application deadline Feb 7): heibrids.berlin/admission/
English
1
12
12
0
Erhan Bas retweetledi
HHMI
HHMI@hhmi_science·
Janelia scientists have mapped the tangled paths of 1,000 neurons in the mouse brain. bit.ly/32FRLhw
English
2
16
40
0
Erhan Bas retweetledi
HHMI
HHMI@hhmi_science·
Scientists have traced the path of more than 80 meters worth of brain cells inside the mouse brain – and they’re just getting started: bit.ly/32ABMRv
GIF
English
0
42
95
0
Erhan Bas retweetledi
HHMI | Janelia
HHMI | Janelia@HHMIJanelia·
What could you achieve in 15 years with $250 million at Janelia? 2-page pre-proposals for our new research area competition are due one month from today. Tell us your big idea! janelia.org/our-research/c…
English
1
49
89
0