Apple removed python 2.7 from macOS Monterey 13.4

Top 8 Blogs for Deep learning you should learn in 2022...✔️✔️✔️

Why you should follow this blog?
👉 Blogs assist you to stay up to date with newly published technology, research, methods, and learning experiences people share with them.

Top 8 Blogs you must read to learn deep learning:
-------------------------------------------------------
1. Towards Data Science: https://lnkd.in/gAB7QSnp

2. Medium.com: https://lnkd.in/gbJaXneh

3. Analytics Vidhya: https://lnkd.in/gk2jevgB

4. KDnuggets: https://lnkd.in/gZUUReq2

5. neptune.ai - https://lnkd.in/gsNksnAD

6. Machine Learning Mastery- https://lnkd.in/gPZft6x3

7. PyImageSearch - https://lnkd.in/gBJUKMby

8. Kaggle - https://lnkd.in/g_3AvFz

9. 100 Blog lists for Data Scientist by Insane: https://lnkd.in/gpsJaveu

Ten HTTP status codes you see most often. 👇🏻

• 200 Ok
Successful request

• 201 Created
New resource was created

• 204 No Content
Successful request but no content

• 304 Not Modified
Use cached version

• 400 Bad Request
Malformed request syntax or message framing

• 401 Unauthorized
Client must authenticate itself

• 403 Forbidden
No access to the content

• 404 Not Found
URL is not recognized

• 429 Too Many Requests
Too many requests in a given timeframe

• 500 Internal Server Error
An unexpected situation that server can't handle

https://www.microsoft.com/en-us/codexchallenge


MSFT code challenge

While looking for an NLP domain to research, I found this amazing resource: https://nlpprogress.com/
It lists out the progress that has been made so far in a variety of different domains so researchers can have a better idea on their areas of improvement before selecting one and diving into a proper literature review -- or just to test out models for their application.

Each section includes the top papers of the domain, the scores achieved in their implementation, and even the code. It's a lot like paperswithcode.com but specifically for NLP.

Debugging NLP models is hard. AdaTest, a new open-source tool, combines the strengths of automated and user-driven methods, leveraging the generative capabilities of large language models and human expertise in an iterative testing and debugging loo

https://github.com/microsoft/adatest

PyTorch Runs On the GPU of Apple M1 Macs Now!

Here’s how long each python package installation took, measuring both wallclock and CPU time:

ML tools - source Springer

https://www.ibm.com/downloads/cas/GB8ZMQZ3

Machine learning Dummies Book to download from IBM

https://github.com/python/peps/blob/3fe4784290a8461dea98efc8fc338662b2992371/pep-0695.rst


PEP 695 proposes a new syntax for declaring type parameters for generic classes, functions, and type aliases.

This will be a big win. The current state of affairs is a recurring source of confusion and is quite unpleasant.

Google AI amazing libs:

TF: 🤖🧠
TFLite: 🤖🧠+📱
TF.js: 🤖🧠+🕸️
JAX: 🧮⚡️
Flax:🤖🧠+🧮⚡️
TF-Agents: 🤖🧠+🕹️♟️
TFX: 🤖🧠+👔🏭
TFHub: 🤖🧠+🧠🧠
TFDF: 🤖🧠+🌳🌲
TensorBoard: 🤖🧠+📊+(🚫🪲)
TF Recommenders: 🤖🧠+(📚➡️📙)
TF Privacy: 🤖🧠+🔐
TF Probability: 🤖🧠+🎲
Magenta: 🤖🧠+🎨🎶

There are three types of transformers models
1. Encoder Only
2. Decode Only
3. Encoder-Decoder

👉 Encoder Only
BERT leverages the Encoder architecture of the transformer. BERT takes text sequence as input. The BERT Encoder produces BERT embedding, which can be used to perform downstream tasks like Text classification or Named Entity Recognition.

👉 Decoder Only
GPT models leverage the decoder architecture of the transformer. Given the input sequence as prompt, GPT starts generating the response. Therefore, GPT models are best suitable for text or sequence generation.

👉 Encoder-Decoder
T5 is a model which uses both Encoder and Decoder architecture. It treats each task as text to text or sequence to sequence. E.g., In text classification, the Encoder takes text as input, and the Decoder generates text labels instead of classifying them.

✍️✍️Latest - Work from home is now a LEGAL RIGHT in Netherlands! 🇳🇱
Dutch employers now MUST consider employee requests to work remotely, as long as their professions allow it...
It's one of the first countries to do so, congrats! 🙌


✍️✍️A new platform launched by Vodafone and Google called AI Booster aims to handle thousands of ML models a day across 18+ countries. AI Booster is the result of 18 months of development and is built upon Google’s Vertex AI and integrates with Vodafone’s Neuron platform.

✍️✍️ Meta, formerly known as Facebook announced that it has built and open-sourced ‘No Language Left Behind’ NLLB-200, a single AI model that is the first to translate across 200 different languages, including 55 African languages with state-of-the-art results.

Last week AI News

✍️✍️Google no longer accepts deepfake projects on Colab
✍️✍️Zoom receives backlash for emotion-detecting AI
✍️✍️ Apple’s former ML director reportedly joins Google DeepMind

10 VScode extensions every data scientist should have💻🤖

1. Python
2. Pylance
3. Python Indent
4. Jupyter
5. Jupyter notebook renderers
6. DVC - (ML model experiment tracking)
7. Gitlens
8. Todo MD
9. Excel viewer
10. Markdown preview GitHub styling

Hugging face released 🌸 BLOOM 🌸 model today, the world’s largest open multilingual language model!

With its 176 billion parameters, BLOOM is able to generate text in 46 natural languages and 13 programming languages.

This release is the culmination of more than a year of work involving over 1000 researchers from 70 countries and 250 institutions, leading to a final run of 11 weeks of training on the Jean Zay supercomputer in the south of Paris, France


Researchers can now download, run and study 🌸 BLOOM 🌸  to investigate the performance and behaviour of recently developed large language models down to their deepest internal operations.

For more information
https://bigscience.huggingface.co/blog/bloom

✍️✍️ 77 per cent of the devices used today have some sort of AI-enabled feature. The global market is projected at USD120 billion by 2025. A post-pandemic world only accelerates that demand. AI is both a challenge and an opportunity.

✍️✍️ Yesterday, admins of the PyPI registry announced they were in the process of introducing two-factor authentication (2FA) requirement for projects deemed "critical."
https://pypi.org/security-key-giveaway/

✍️✍️ AI bot wrote a scientific paper on itself 2-hours.

✍️✍️ DeepMind AI learns simple physics like a baby. The software model, named Physics Learning through Auto-encoding and Tracking Objects (PLATO), was fed the raw images from the videos, but also versions that highlighted each object in the scene. PLATO was designed to develop an internal representation of the physical properties of the objects, such as their positions and velocities.

✍️✍️ Google and Microsoft call ‘EMOTION AI’ risky but only limit usage.


✍️✍️ Indian Defence Minister launched 75 defence products powered by AI on July 11th, Monday. These products will be found at the first-ever ‘Artificial Intelligence in Defence’ (AIDef) symposium.

✍️✍️ Indian Railways to install AI-enabled CCTV cameras by Jan 2023. All these CCTV cameras under the VSS(Video Surveillance System) at railway stations, which are major hubs of transportation, would be Internet Protocol (IP) covering the waiting halls, reservation counters, parking areas, main entrance and exit, platforms, foot over bridges, booking offices", said the railway


Failure Usecase in AI
✍️✍️ In 2013, the Dutch government deployed artificial intelligence to handle childcare benefits applications and, as you might guess, it did not go well. Disproportionately, ethnic minorities were denied benefits and charged with fraud, and the entire imbroglio culminated in the cabinet resigning in January 2021. Now, it seems, it may not have been a fault of technology as much as a fault of the human beings, or policymakers, operating the technology.