TensorFlow, Keras and deep learning, without a PhD access_tim

https://codelabs.developers.google.com/codelabs/cloud-tensorflow-mnist/#2

Learning to Combine Top-Down and Bottom-Up Signals in Recurrent Neural Networks with Attention over Modules Mittal et al.: #ArtificialIntelligence #DeepLearning #MachineLearning

https://arxiv.org/abs/2006.16981

How to Write a Makefile - Automating Python Setup, Compilation, and Testing

https://stackabuse.com/how-to-write-a-makefile-automating-python-setup-compilation-and-testing/

How to read Deep Learning research papers.

⚫ A systematic approach to reading a collection of papers to gain knowledge within a domain
⚫ How to properly read a research paper
⚫ Useful online resources that
can aid you in searching for papers and key information "50–100 papers will primarily provide you with a very good understanding of the domain."

https://towardsdatascience.com/how-you-should-read-research-papers-according-to-andrew-ng-stanford-deep-learning-lectures-98ecbd3ccfb3

All the videos for the Computer Vision lecture
on "Detection, Segmentation, and Tracking" are now public!

Videos: https://youtube.com/playlist?list=PLog3nOPCjKBneGyffEktlXXMfv1OtKmCs…
Slides: https://dvl.in.tum.de/teaching/cv3dst-ss20/

In future #AI hiring other AI be like: Job Profile: *human baby sitter*
- Experience : trained on 100 years of past data.
- Test Accuracy : 99.9999
- Precision: blah
- recall : blah
- AUC : blah blah
- Inference time: A.C
- Trained on : Latest "alien" TPUs and GPUs
- Bias : blah Note: AI trained on old TPUs will not be considered. And then AI will gossip with each other about bias and discrimination they have to go through compared to others like:
- "Wouldn't I be considered if I am trained on X country's data?"
- "Why was she considered even though she has outliers in the data?"
- "I am trained on old TPUs, I won't be considered? What!" LOL #artificialintelligence #machinelearning

A Hybrid Approach for Fake News Detection in Twitter Based on User Features and Graph Embeddings

• Using node2vec to extract features from a twitter follower graph. In conjunction with user features provided by Twitter.

This hybrid approach considers both the characteristics of the user and his social graph. The results show that the approach consistently and significantly outperforms existent approaches limited to user features.

Paper is.gd/LP9uKD

Stanford CS224w’s lectures Machine Learning with Graphs, Leskovec et al.: https://lnkd.in/d4Cnahj #DeepLearning #Graphs #MachineLearning

Locality sensitive hashing:
https://medium.com/carbon-consulting/explaining-lsh-minhash-simhash-c3cc33040030

❇️ @AI_Python
🗣 @AI_Python_arXiv
✴️ @AI_Python_en

Tensorflow + dalex, or how to explain a TensorFlow model

Article:
https://www.kdnuggets.com/2020/11/dalex-explain-tensorflow-model.html

Code:
https://dalex.drwhy.ai/python-dalex-tensorflow.html

Explanatory Model Analysis:
https://pbiecek.github.io/ema/

The upcoming Saturday 10am CET we perform a SeaMapAI Project Meetup #1. The idea is to cycle GANs, utilize neural networks orchestration, evaluate images and eventually generate sea maps!

https://www.eventbrite.co.uk/e/introducing-sailmapai-educational-neural-seamap-generator-project-tickets-149691925579?aff=telegram

If you are looking for educational opportunities, subscribe to this channel
👇👇
🔰 https://t.me/DLeX_Apply

If there is no signal in your data, the ML model won't magically be predictive.

Plus simpler models will do better with low signal vs. bigger more complex models...

Unfortunately, it is, what it is

he more you train deep learning models, the more you realize that there is still so much left to figure out with NNs..

Neural networks still

- require a lot of data cleaning
- need a fair bit of featurization
- regularly overfit
- often don't learn the nuance in the data

Self supervised learning is the most intriguing form of AI yet..

Like babies, machine simply learn by observing the environment

Multi-task self learners that learn from hybrid inputs comprising of text, voice and images will be our next step towards AGI

با تشکر از دکتر آرش ربانی برای ارسال این موقعیت تحصیلی هر کدوم از دانشجوها علاقمند به اپلای هستند اقدام کنند.
A fully funded Ph.D. position in computer science at the University of Leeds! We are looking for an outstanding international or UK-based applicant to join our research group (Data Flow Lab) at the school of computing. The project title is “3D reconstruction of porous material based on 2D images using conditional generative adversarial neural networks (CGANs)”. If you are interested in Generative Adversarial Neural Networks, functional material, porous media, and 3D image analysis, don’t miss the opportunity! We will be working closely with the school of geoscience and medicine to study realistic images of geological and biological porous material. For more information and to express interest, please be in touch through email (a.rabbani@leeds.ac.uk).

FindAPhD page link: https://www.findaphd.com/phds/project/3d-reconstruction-of-porous-material-based-on-2d-images-using-conditional-generative-adversarial-neural-networks-cgans/?p154984
University page link: https://phd.leeds.ac.uk/project/1597-3d-reconstruction-of-porous-material-based-on-2d-images-using-conditional-generative-adversarial-neural-networks-cgans
Research group link: www.DataFlowLab.org

Introduction to Deep Learning, Carnegie Mellon 2022-23

https://m.youtube.com/playlist?list=PLp-0K3kfddPwgBSCbDtT6NaVOd-gIHVMW

محدودیتهای مدلهای LLMs

Fundamental Limitations of Alignment in Large Language Models

https://arxiv.org/abs/2304.11082

#مقاله #ایده_جذاب

🔸 مطالب بیشتر 👇👇

✅ @AI_DeepMind