اگر با کراس کد زده باشید و اندازه کار بزرگ بوده باشه میدونید که یک سری محدودیت‌های خسته کننده برای
distributed training
وجود داره، البته که می‌شه مشکلات رو حل کرد اما خب زمانبر هست

توی تنسورفلو یک، ،submodule دیگری هم وجود داره به اسم estimator که توی نسخه ۲ توجه خوبی بهش شده و پیشرفتای خوبی داشته، حتی خیلی ریکامند می‌شه که موقع
Distributed Processing
بجای کراس ازین مورد استفاده بشه، ولی چون نوع syntax خودش رو داره برای کدهای بزرگ که روی کراس نوشته شده بنظر مفید نمیاد
اما :
tf.keras.estimator.model_to_estimator()
مشکل رو حل می‌کنه؛ توی تست‌های بنده اگر از خود کراس مستقیم استفاده کنید
pip install keras
و بخواید اینکارو انجام بدید مشکلاتی پیش میاد اما نسخه تنسورفلو ۲ به راحتی و عالی اینکارو انجام میده، دوم اینکه دیگه نیازی نیس یادتون بمونه که حتماً از
@tf.function
استفاده کنید، چون این مورد خوردش بهینه‌ساز‌ی‌هارو انجام میده، پردازش توزیع شده هم که دلیل اصلی استفاده هست

Statistical Modelling vs Machine Learning

At times it may seem Machine Learning can be done these days without a sound statistical background but those people are not really understanding the different nuances. Code written to make it easier does not negate the need for an in-depth understanding of the problem.

Link

🔭 @DeepGravity

Self-Tuning Deep Reinforcement Learning

Reinforcement learning (RL) algorithms often require expensive manual or automated hyperparameter searches in order to perform well on a new domain. This need is particularly acute in modern deep RL architectures which often incorporate many modules and multiple loss functions. In this paper, we take a step towards addressing this issue by using metagradients (Xu et al., 2018) to tune these hyperparameters via differentiable cross validation, whilst the agent interacts with and learns from the environment. We present the Self-Tuning Actor Critic (STAC) which uses this process to tune the hyperparameters of the usual loss function of the IMPALA actor critic agent(Espeholt et. al., 2018), to learn the hyperparameters that define auxiliary loss functions, and to balance trade offs in off policy learning by introducing and adapting the hyperparameters of a novel leaky V-trace operator. The method is simple to use, sample efficient and does not require significant increase in compute. Ablative studies show that the overall performance of STAC improves as we adapt more hyperparameters. When applied to 57 games on the Atari 2600 environment over 200 million frames our algorithm improves the median human normalized score of the baseline from 243

Paper

🔭 @DeepGravity

Attention Augmented Convolutional Networks

Convolutional neural networks have proven to be a powerful tool for image recognition, allowing for ever-improving results in image classification (ImageNet), object detection (COCO), and other tasks. Despite their success, convolutions are limited by their locality, i.e. their inability to consider relations between different areas of an image. On the other hand, a popular mechanism which has proven success in overcoming locality is self-attention, which has shown to be able to capture long-range interactions (e.g. Show, Attend and Tell).

Article

🔭 @DeepGravity

#probability and Coronavirus

From #statistical point of view, there is an important difference between having any sort of direct or indirect interaction with a person with #coronavirus and risk of getting the disease. For example, people sitting in window seats have lower chance of getting the disease if there would be a person with the virus in a flight (check the figure). Or if the person with the disease walk and have direct or indirect interaction with other people, it is mainly the people in two seats around him that have the high probability of getting the disease.

These are just some statistical facts but anyway we have to be very cautious.
Note. Don’t avoid interaction with people based on race.

Credit goes to Seyed Ali Madani

🔭 @DeepGravity

Google's DeepMind just shared AI-generated predictions about the #coronavirus that could help researchers stem the global outbreak

Link

🔭 @DeepGravity

Exponential growth and epidemics

#coronavirus

YouTube

🔭 @DeepGravity

RCC-Dual-GAN: An Efficient Approach for Outlier Detection with Few Identified Anomalies

Outlier detection is an important task in data mining and many technologies have been explored in various applications. However, due to the default assumption that outliers are non-concentrated, unsupervised outlier detection may not correctly detect group anomalies with higher density levels. As for the supervised outlier detection, although high detection rates and optimal parameters can usually be achieved, obtaining sufficient and correct labels is a time-consuming task. To address these issues, we focus on semi-supervised outlier detection with few identified anomalies, in the hope of using limited labels to achieve high detection accuracy. First, we propose a novel detection model Dual-GAN, which can directly utilize the potential information in identified anomalies to detect discrete outliers and partially identified group anomalies simultaneously. And then, considering the instances with similar output values may not all be similar in a complex data structure, we replace the two MO-GAN components in Dual-GAN with the combination of RCC and M-GAN (RCC-Dual-GAN). In addition, to deal with the evaluation of Nash equilibrium and the selection of optimal model, two evaluation indicators are created and introduced into the two models to make the detection process more intelligent. Extensive experiments on both benchmark datasets and two practical tasks demonstrate that our proposed approaches (i.e., Dual-GAN and RCC-Dual-GAN) can significantly improve the accuracy of outlier detection even with only a few identified anomalies. Moreover, compared with the two MO-GAN components in Dual-GAN, the network structure combining RCC and M-GAN has greater stability in various situations.

Paper

🔭 @DeepGravity

Advancements in Reinforcement Learning with Sergey Levine

Podcast

🔭 @DeepGravity

Variational Methods for Computer Vision Course

Summary
Variational Methods are among the most classical techniques for optimization of cost functions in higher dimension. Many challenges in Computer Vision and in other domains of research can be formulated as variational methods. Examples include denoising, deblurring, image segmentation, tracking, optical flow estimation, depth estimation from stereo images or 3D reconstruction from multiple views.

In this class, I will introduce the basic concepts of variational methods, the Euler-Lagrange calculus and partial differential equations. I will discuss how respective computer vision and image analysis challenges can be cast as variational problems and how they can be efficiently solved. Towards the end of the class, I will discuss convex formulations and convex relaxations which allow to compute optimal or near-optimal solutions in the variational setting.

Course by Technical University of Munich

🔭 @DeepGravity

Zoom In: An Introduction to Circuits
By studying the connections between neurons, we can find meaningful algorithms in the weights of neural networks.

Article and Codes

🔭 @DeepGravity

#Covid19, your community, and you — a data science perspective

We are data scientists—that is, our job is to understand how to analyze and interpret data. When we analyze the data around covid-19, we are very concerned. The most vulnerable parts of society, the elderly and the poor, are most at risk, but controlling the spread and impact of the disease requires us all to change our behavior. Wash your hands thoroughly and regularly, avoid groups and crowds, cancel events, and don’t touch your face. In this post, we explain why we are concerned, and you should be too. For an excellent summary of the key information you need to know, read Corona in Brief by Ethan Alley (the president of a non-profit that develops technologies to reduce risks from pandemics).

Link to the community

🔭 @DeepGravity

Generative Teaching Networks: Accelerating Neural Architecture Search by Learning to Generate Synthetic Training Data

Article

🔭 @DeepGravity

Deep Learning System to Screen #Coronavirus Disease 2019 Pneumonia

We found that the real time reverse trannoscription-polymerase chain reaction (RT-PCR) detection of viral RNA from sputum or nasopharyngeal swab has a relatively low positive rate in the early stage to determine COVID-19 (named by the World Health Organization). The manifestations of computed tomography (CT) imaging of COVID-19 had their own characteristics, which are different from other types of viral pneumonia, such as Influenza-A viral pneumonia. Therefore, clinical doctors call for another early diagnostic criteria for this new type of pneumonia as soon as possible.This study aimed to establish an early screening model to distinguish COVID-19 pneumonia from Influenza-A viral pneumonia and healthy cases with pulmonary CT images using deep learning techniques. The candidate infection regions were first segmented out using a 3-dimensional deep learning model from pulmonary CT image set. These separated images were then categorized into COVID-19, Influenza-A viral pneumonia and irrelevant to infection groups, together with the corresponding confidence scores using a location-attention classification model. Finally the infection type and total confidence score of this CT case were calculated with Noisy-or Bayesian function.The experiments result of benchmark dataset showed that the overall accuracy was 86.7 % from the perspective of CT cases as a whole.The deep learning models established in this study were effective for the early screening of COVID-19 patients and demonstrated to be a promising supplementary diagnostic method for frontline clinical doctors.

Paper

🔭 @DeepGravity

On the Robustness of Cooperative Multi-Agent Reinforcement Learning

In cooperative multi-agent reinforcement learning (c-MARL), agents learn to cooperatively take actions as a team to maximize a total team reward. We analyze the robustness of c-MARL to adversaries capable of attacking one of the agents on a team. Through the ability to manipulate this agent's observations, the adversary seeks to decrease the total team reward. Attacking c-MARL is challenging for three reasons: first, it is difficult to estimate team rewards or how they are impacted by an agent mispredicting; second, models are non-differentiable; and third, the feature space is low-dimensional. Thus, we introduce a novel attack. The attacker first trains a policy network with reinforcement learning to find a wrong action it should encourage the victim agent to take. Then, the adversary uses targeted adversarial examples to force the victim to take this action. Our results on the StartCraft II multi-agent benchmark demonstrate that c-MARL teams are highly vulnerable to perturbations applied to one of their agent's observations. By attacking a single agent, our attack method has highly negative impact on the overall team reward, reducing it from 20 to 9.4. This results in the team's winning rate to go down from 98.9

Paper

🔭 @DeepGravity

سلام دوستان
من یه رپوزیتوری اولیه آماده کردم که بتونیم اونجا تمامی داده‌های مربوط به کرونای ایران رو جمع کنیم تا بشه اونا رو آنالیز هم کرد.

از دوستانی که علاقه‌مند هستند کمک کنند لطفا به من ایمیل بزنند. باید ابتدا خود رپو رو آماده کنیم و لایسنس‌های لازم رو بهش اد کنیم. سپس همه‌ی عزیزانی که در ایران هستن و به داده‌ها دسترسی دارن لطفا داده‌ها رو به رپو اد کنن تا بتونیم آنالیزشون کنیم. طبیعتا با حفظ نکات امنیتی مربوط به داده‌ها.
احتمالا باید پرتوکول‌هایی برای این کار طراحی کنیم، ولی من خودم ممکنه دقیق همه رو ندونم. بنابراین دوستانی که می‌تونن کمک کنند لطفا به من اطلاع بدن:
DeepL.Gravity@gmail.com

با سپاس و با این امید که ما هم بتونیم کمک کرده باشیم.
ارادتمند

#Covid19

🔭 @DeepGravity

Detecting #COVID19 in X-ray images with Keras, TensorFlow, and Deep Learning

Link

🔭 @DeepGravity

Mehdi Allahyari is an Assistant Professor of Computer Science at Georgia Southern University

Here you might find his Persian course in Neural Net

🔭 @DeepGravity

درود بر همه‌ی شما دوستان گرامی،
امیدوارم این روزهای سخت بهاری به زودی با چیرگی سبزی بر سیاهی سپری بشه. هر چند اندوهش هرگز از یادها نخواهد رفت.

به منظور بررسی ابعاد بحران #کرونا از نگاه #ماشین_لرنینگ، قصد دارم به کمک شما عزیزان یک جلسه‌ی هم‌اندیشی آنلاین رو راه‌اندازی کنم. در لینک زیر زمان‌های مختلفی رو می‌بینین. لطفا زمانی که برای شما مناسب‌تره رو انتخاب کنین که تو اون تایم از طریق زوم یا گوگل میت بتونیم دور هم جمع بشیم. سعی کردم گزینه‌ها رو بین صبح و عصر و شب پخش کنم که با توجه به اختلاف ساعت‌ها بتونیم تایم مشترکی رو پیدا کنیم:

https://doodle.com/poll/69fvgkegwq3y8p6w

هدف این جلسه بیشتر هم اندیشی و به اشتراک گذاری دانسته‌ها و داشته‌ها ست. من خودم دو تا رپو آماده کردم که در موردشون توضیح خواهم داد.
(هدف مقاله دادن یا کار اقتصادی کردن نیست)

امیدوارم ما هم بتونیم در کنار تیم درمان، کمکی برای کشور (و شاید دنیا) در این شرایط باشیم.

اگه پیشنهادی هم دارین، لطفا در کامنت یا به صورت خصوصی پیام بذارین.

ارادتمند
#ai #computervision #machinelearning #deeplearning #covid19

@Reza

🔭 @DeepGravity

سلام برهمگی. عزیزانی که علاقه مند هستن در این هم اندیشی شرکت کنن، طبق رای گیری انجام شده، هم اندیشی در تاریخ 14 فروردین ساعت 11:30 به وقت ایران و در میتینگ روم زیر برگزار میشه.



https://hslu.zoom.us/j/6364857897