Large teams develop and small teams disrupt science and technology. Truly transformative research is done by small teams.
π Link
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
π Link
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
Vidhyalakshmi Sundara Raman
Let's say you open-source your MNIST-trained ConvNet and the Postal Service decides to use it to read zipcodes.
Since your ConvNet is open source, someone could make a malicious white-box attack that would imperceptibly modify every envelope so as to maximally confuse your ConvNet.
Swarms of painter-drones would attack mailboxes and carefully corrupt every zipcode.
Every piece of junk mail would end up in the wrong zipcode!
Surely, civilization would crumble as a result.
If I were you, I would buy a place on a remote island to survive the inevitable armageddon, like, right now.
#regulateMNIST
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
Let's say you open-source your MNIST-trained ConvNet and the Postal Service decides to use it to read zipcodes.
Since your ConvNet is open source, someone could make a malicious white-box attack that would imperceptibly modify every envelope so as to maximally confuse your ConvNet.
Swarms of painter-drones would attack mailboxes and carefully corrupt every zipcode.
Every piece of junk mail would end up in the wrong zipcode!
Surely, civilization would crumble as a result.
If I were you, I would buy a place on a remote island to survive the inevitable armageddon, like, right now.
#regulateMNIST
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
HOW CMLD TRANSFORMING MACHINE LEARNING PROFESSIONSHOW CMLD TRANSFORMING MACHINE LEARNING PROFESSIONS
https://www.youtube.com/watch?v=OrQqtC1YSRI
IF INTERESTED https://www.bepec.in/cmld
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
https://www.youtube.com/watch?v=OrQqtC1YSRI
IF INTERESTED https://www.bepec.in/cmld
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
Techniques for Detection of Rusting of Metals using Image Processing: A Survey
Most of industries around us make use of iron machines & tools for manufacturing their products. On the other
hand corrosion is a natural process that deteriorates the integrity
of iron surface. Therefore, rusting of iron takes place. To avoid unwanted accidents in industries, it is necessary to detect rusting in earlier stage, so that it can be prevented. Digital Image Processing for the detection of the rusting provides fast, accurate and objectives results. There have been many techniques for detection of rust. In this paper we are describing some existing techniques for detection of rusting. We have analyzed these
techniques and made comparison based o their approaches,
strengths and limitations.
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.675.7088&rep=rep1&type=pdf
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
Most of industries around us make use of iron machines & tools for manufacturing their products. On the other
hand corrosion is a natural process that deteriorates the integrity
of iron surface. Therefore, rusting of iron takes place. To avoid unwanted accidents in industries, it is necessary to detect rusting in earlier stage, so that it can be prevented. Digital Image Processing for the detection of the rusting provides fast, accurate and objectives results. There have been many techniques for detection of rust. In this paper we are describing some existing techniques for detection of rusting. We have analyzed these
techniques and made comparison based o their approaches,
strengths and limitations.
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.675.7088&rep=rep1&type=pdf
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
Highly recommend MIT 6.S191: Introduction to Deep Learning Course by Alexander Amini and Ava Soleimany through Massachusetts Institute of Technology (MIT)
#artificialintelligence
#deeplearning
#machinelearning
#CNN
#vision
MIT Deep Learning Course 6.S191 https://www.linkedin.com/pulse/mit-deep-learning-course-6s191-bhagirath-kumar-lader
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
#artificialintelligence
#deeplearning
#machinelearning
#CNN
#vision
MIT Deep Learning Course 6.S191 https://www.linkedin.com/pulse/mit-deep-learning-course-6s191-bhagirath-kumar-lader
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
Rust Detection Using #DeepLearning
https://www.youtube.com/watch?v=iYsSmnma8jA
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
https://www.youtube.com/watch?v=iYsSmnma8jA
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
I didn't try a wide range of libraries related to NLP but SpaCy is my best one so far. It is written in #Cython and it is 100 times faster than NLTK, and very Pythonic.
If you are familiar with the Python data science stack, spaCy is your Numpy for #NLP.
Code examples:
( https://lnkd.in/dXhzbJt )
( https://lnkd.in/dYtArJ7 )
No alt text provided for this image
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
If you are familiar with the Python data science stack, spaCy is your Numpy for #NLP.
Code examples:
( https://lnkd.in/dXhzbJt )
( https://lnkd.in/dYtArJ7 )
No alt text provided for this image
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
NNI (Neural Network Intelligence) is a toolkit to help users run automated machine learning (AutoML) experiments. #automation #machinelearning parameters
https://github.com/Microsoft/nni
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
https://github.com/Microsoft/nni
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
With the plethora of Deep learning courses available on the internet, this (CS 598 LAZ: Cutting-Edge Trends in Deep Learning and Recognition) stands out as one of the best I have seen.
For instance, there 274 slides on meta-learning. (https://lnkd.in/enuZhAX).
https://lnkd.in/ehFAQew
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
For instance, there 274 slides on meta-learning. (https://lnkd.in/enuZhAX).
https://lnkd.in/ehFAQew
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
This person does not exist
By Philip Wang: https://lnkd.in/eEWxyYu
#GenerativeAdversarialNetworks #GAN
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
By Philip Wang: https://lnkd.in/eEWxyYu
#GenerativeAdversarialNetworks #GAN
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
Media is too big
VIEW IN TELEGRAM
Introducing a new blog (0:00-0:50) and then explaining what transfer learning is and why people have the wrong idea about it (0:50-03:05).
It's my first one of these, so go easy on me! <3
Cassie's blog: https://lnkd.in/e55m6sY (Start here: https://lnkd.in/eE4Vg6r)
Completely new to machine learning? Here's my intro: https://lnkd.in/eVuKzNe
Launchpad blog: https://lnkd.in/e3EW89u
Keeping up with AI: https://lnkd.in/ePqWvC3
View on YouTube: https://lnkd.in/eExmdET
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
It's my first one of these, so go easy on me! <3
Cassie's blog: https://lnkd.in/e55m6sY (Start here: https://lnkd.in/eE4Vg6r)
Completely new to machine learning? Here's my intro: https://lnkd.in/eVuKzNe
Launchpad blog: https://lnkd.in/e3EW89u
Keeping up with AI: https://lnkd.in/ePqWvC3
View on YouTube: https://lnkd.in/eExmdET
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
Quantum Machine Learning is not Machine Learning for Quantum Mechanics
There is a bit of confusion in the literature and research. Applying machine learning to resolve numerically accurate quantum mechanics applied to molecules does not constitute 'quantum machine learning' (QML). QML implies machine learning algorithms solved within a quantum mechanical computing device. Common misconception. Recent work does this error though work is very original applying Kernel methods in response properties.
#ai #quantumcomputing #machinelearning #datascience
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
There is a bit of confusion in the literature and research. Applying machine learning to resolve numerically accurate quantum mechanics applied to molecules does not constitute 'quantum machine learning' (QML). QML implies machine learning algorithms solved within a quantum mechanical computing device. Common misconception. Recent work does this error though work is very original applying Kernel methods in response properties.
#ai #quantumcomputing #machinelearning #datascience
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
Forwarded from Ψ―Ψ³ΨͺΨ§ΩΨ±Ψ―ΩΨ§Ϋ ΫΨ§Ψ―Ϊ―ΫΨ±Ϋ ΨΉΩ
ΫΩ(InTec)
Hi
Highly recommended to see what is going on in these scientific Links ( Code, articles, summary , discussions .... )
AI Articles with Code ( most in Python )
https://www.paperswithcode.com/
Public summaries in Machine learning , organized with community ( Like stackoverflowin AI )
https://www.shortscience.org/?s=cs
Search /Filtering recent Arxiv ( pre print submissions ) ; Keep track of recent papers with sorting papers by similarity
http://www.arxiv-sanity.com
Latest Arxiv papers with abstract summary ! ( web interfece and python code )
https://github.com/chiphuyen/sotawhat
More info here :
https://huyenchip.com/2018/10/04/sotawhat.html
ΨͺΨ΄Ϊ©Ψ± ΩΫΪΩ Ψ§Ψ² Ψ―Ϊ©ΨͺΨ± ΨΨ¨ΫΨ¨βΨ²Ψ§Ψ―Ω Ψ¨Ψ§Ψ¨Ψͺ Ψ¨Ω Ψ§Ψ΄ΨͺΨ±Ψ§Ϊ© Ϊ―Ψ°Ψ§Ψ±Ϋ Ω ΩΨ§Ψ¨ΨΉ
@pytens
Highly recommended to see what is going on in these scientific Links ( Code, articles, summary , discussions .... )
AI Articles with Code ( most in Python )
https://www.paperswithcode.com/
Public summaries in Machine learning , organized with community ( Like stackoverflowin AI )
https://www.shortscience.org/?s=cs
Search /Filtering recent Arxiv ( pre print submissions ) ; Keep track of recent papers with sorting papers by similarity
http://www.arxiv-sanity.com
Latest Arxiv papers with abstract summary ! ( web interfece and python code )
https://github.com/chiphuyen/sotawhat
More info here :
https://huyenchip.com/2018/10/04/sotawhat.html
ΨͺΨ΄Ϊ©Ψ± ΩΫΪΩ Ψ§Ψ² Ψ―Ϊ©ΨͺΨ± ΨΨ¨ΫΨ¨βΨ²Ψ§Ψ―Ω Ψ¨Ψ§Ψ¨Ψͺ Ψ¨Ω Ψ§Ψ΄ΨͺΨ±Ψ§Ϊ© Ϊ―Ψ°Ψ§Ψ±Ϋ Ω ΩΨ§Ψ¨ΨΉ
@pytens
Paperswithcode
Papers with Code - The latest in Machine Learning
Papers With Code highlights trending Machine Learning research and the code to implement it.
OpenAI built a text generator so good, itβs considered too dangerous to release https://buff.ly/2S61M1v #AI #ArtificialIntelligence #MachineLearning
How to Design a Neural Network for Image Classification
Here is a simple illustration of what a shallow and deep neural network looks like.
#DeepLearning #Fundamentals #neuralnetworks #design
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
Here is a simple illustration of what a shallow and deep neural network looks like.
#DeepLearning #Fundamentals #neuralnetworks #design
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
The AI Now Institute has published a report outlining the challenges of law enforcement using AI algorithms to help forecast criminal activity.
The research center based at New York University focuses on the social impact of AI. The paper shows the negative effects of relying on flawed data and focuses on thirteen case studies from different law enforcement agencies in the US.
For example, βdirty dataβ contains hidden biases that might predict that certain areas have elevated levels of crime. More police may be deployed in that area, leading to more racial profiling and arrests.
https://lnkd.in/dtx_2rg
#deeplearning #machinelearning #police #lawenforcement #algorithms
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
The research center based at New York University focuses on the social impact of AI. The paper shows the negative effects of relying on flawed data and focuses on thirteen case studies from different law enforcement agencies in the US.
For example, βdirty dataβ contains hidden biases that might predict that certain areas have elevated levels of crime. More police may be deployed in that area, leading to more racial profiling and arrests.
https://lnkd.in/dtx_2rg
#deeplearning #machinelearning #police #lawenforcement #algorithms
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
There is the mathematical aspect to statistics - an important one - that can turn a lot of "normal" people off. But there is also applied statistics which, in John Tukey's words, lets us play in everyone's backyard.
Everyone, in this case, includes psychology, sociology, history, anthropology, physics, biology, economics, political science, and even art, literature, music and philosophy. Stats opens doors.
It can help us better understand the workings of the human mind, how societies and cultures function, how to interpret historical facts meaningfully and how think about the future.
One can learn a lot about human nature by studying a jobs or educational program, for example.
Most fundamentally perhaps, it shows us better ways to think about causation.
If a statistician only focuses on the math and programing, though, they will miss all of this. It seems a shame, but different strokes for different folks.
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
Everyone, in this case, includes psychology, sociology, history, anthropology, physics, biology, economics, political science, and even art, literature, music and philosophy. Stats opens doors.
It can help us better understand the workings of the human mind, how societies and cultures function, how to interpret historical facts meaningfully and how think about the future.
One can learn a lot about human nature by studying a jobs or educational program, for example.
Most fundamentally perhaps, it shows us better ways to think about causation.
If a statistician only focuses on the math and programing, though, they will miss all of this. It seems a shame, but different strokes for different folks.
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
MIT : Intro to Deep Learning
First two 2019 lectures for MIT Intro to #DeepLearning now online!
Course schedule: https://lnkd.in/eDW7FTs
Lecture 1: https://lnkd.in/esDcMaP
Lecture 2: https://lnkd.in/epzKtXM
#artificialinteligence #machineleaning #neuralnetworks
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
First two 2019 lectures for MIT Intro to #DeepLearning now online!
Course schedule: https://lnkd.in/eDW7FTs
Lecture 1: https://lnkd.in/esDcMaP
Lecture 2: https://lnkd.in/epzKtXM
#artificialinteligence #machineleaning #neuralnetworks
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
Accenture's 10 Essential ML Interview Questions (with Answers) by The Learning Machine!
https://www.thelearningmachine.ai/accenture
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
https://www.thelearningmachine.ai/accenture
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
image_2019-02-20_12-48-38.png
872.8 KB
Deep Convolutional Sum-Product Networks for Probabilistic Image Representations
Sum-Product Networks (SPNs) are hierarchical probabilistic graphical models capable of fast and exact inference.
Applications of SPNs to real-world data such as large image datasets has been fairly limited in previous literature. Here is a Convolutional Sum-Product Networks (ConvSPNs) which exploit the inherent structure of images in a way similar to deep convolutional neural networks, optionally with weight sharing.
#neuralnetworks #datasets #deeplearning
Paper: https://lnkd.in/ei4Gqjy
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv
Sum-Product Networks (SPNs) are hierarchical probabilistic graphical models capable of fast and exact inference.
Applications of SPNs to real-world data such as large image datasets has been fairly limited in previous literature. Here is a Convolutional Sum-Product Networks (ConvSPNs) which exploit the inherent structure of images in a way similar to deep convolutional neural networks, optionally with weight sharing.
#neuralnetworks #datasets #deeplearning
Paper: https://lnkd.in/ei4Gqjy
β΄οΈ @AI_Python_EN
βοΈ @AI_Python
π£ @AI_Python_arXiv