انسان در واقع اشیاء را بدون ناظر یاد می‌گیرد و بعد اینکه مثلا مدتی یک شی را دید و یاد گرفت، بلافاصله پس از اینکه نام آن شی را شنید برچسب آن را نیز یاد میگیرد.
در حال حاضر بهترین مدلهای بینایی ماشین که در سالهای اخیر، خصوصا بعد از الکس‌نت سال 2012 ارائه شده اند با ناظر هستند. خیلی خوب عمل می‌کنند اما به داده ی برچسب گذاری شده ی زیادی نیاز دارند.
اگر به نحوی بتوانیم از داده های بدون برچسب استفاده کنیم و مدل را آموزش دهیم، سپس در فاز کوتاهی با داده های اندک اشیائی که مدل یاد گرفته است را به صورت با ناظرآموزش دهیم تحول بزرگی در یادگیری مدل ها ایجاد خواهد شد. در این صورت میتوان به سادگی میلیون ها ساعت ویدیو را مثلا با استفاده از یوتیوب به مدل آموزش داد و پس از آموزش مدل، شروع به آموزش نام اشیاء یادگرفته شده به مدل پرداخت روندی که در انسان هم مشاهده میشود! در واقع کودک از بدو تولد اشیاء مختلف را میبیند و آن ها را یاد میگیرد اما با یک یا چند بارشنیدن نام آن به آن دسته یا شئی که قبلا فراگرفته نام اختصاص میدهد.

The Next Frontier in AI: Unsupervised Learning
#Yann_LeCun
Director of AI Research at Facebook, Professor of Computer Science, New York University

November 18, 2016


https://www.youtube.com/watch?v=IbjF5VjniVE

Abstract
The rapid progress of #AI in the last few years are largely the result of advances in #deep_learning and neural nets, combined with the availability of large datasets and fast GPUs. We now have systems that can #recognize images with an accuracy that rivals that of humans. This will lead to revolutions in several domains such as autonomous transportation and #medical #image understanding. But all of these systems currently use #supervised learning in which the machine is trained with inputs labeled by humans. The challenge of the next several years is to let machines learn from raw, #unlabeled_data, such as #video or #text. This is known as #unsupervised learning. AI systems today do not possess "common sense", which humans and animals acquire by observing the world, acting in it, and understanding the physical constraints of it. Some of us see unsupervised learning as the key towards machines with common sense. Approaches to unsupervised learning will be reviewed. This presentation assumes some familiarity with the basic concepts of deep learning.

مقاله‌ی جالب کمپانی اپل!
( Submitted for review to a conference on Nov 15, 2016)

✏️Title:
Learning from Simulated and Unsupervised Images through Adversarial Training

✏️abstract:
With recent progress in graphics, it has become more tractable to train models on #synthetic images, potentially avoiding the need for expensive annotations. However, learning from synthetic images may not achieve the desired performance due to a gap between synthetic and real image distributions. To reduce this gap, we propose Simulated+Unsupervised (S+U) learning, where the task is to learn a model to improve the realism of a simulator's output using #unlabeled real data, while preserving the annotation information from the simulator. We develop a method for S+U learning that uses an #adversarial network similar to #Generative Adversarial Networks (#GANs), but with synthetic images as inputs instead of random vectors. We make several key modifications to the standard GAN algorithm to preserve annotations, avoid artifacts and stabilize training: (i) a 'self-regularization' term, (ii) a local adversarial loss, and (iii) updating the discriminator using a history of refined images. We show that this enables generation of highly realistic images, which we demonstrate both qualitatively and with a user study. We quantitatively evaluate the generated images by training models for gaze estimation and hand pose estimation. We show a significant improvement over using synthetic images, and achieve state-of-the-art results on the MPIIGaze dataset without any labeled real data.

🔗https://arxiv.org/abs/1612.07828v1
🔗https://arxiv.org/pdf/1612.07828v1.pdf

#unlabeled_data #unsupervised #unsupervised_learning #Generative #Generative_Models