This is the Ideal Place to Start Buying after the Market decline

Let the Stage1 form and then buy with patience when it moves to Stage2

Cash is a position 👌

Buy the Dip is not required to get Rich

#VVVLearning

#Thread - 1/21

A story of a brave woman who let her "breast" cut to protect the life of Netaji Subhash Chandra Bose!

Neera Arya was born on March 5 1902 in Khekra Nagar, Uttar Pradesh in a distinguished & eminent businessman Seth Chhajumal's family. Since her childhood,

2/21
She was a true nationalist & always had a vision of becoming a part of the freedom fighter movement. She was also reportedly accused as an undercover by the British Govt.

That was the time when her father's business was flourishing in Calcutta, though his business was..

3/21
spread across the country but Calcutta was the center of his business and that was one of the reasons why her education was done in Calcutta. During that period of time, she had learned quite a few languages along with Bengali, Hindi, English.

4/21
Her father got her married to a CID Inspector "Shrikant Jairanjan Das" in British India but they were two different individuals, Neera Arya was a true nationalist but her husband was a true British servant. She was highly motivated & she had an urge of getting us freedom..

5/21
so she ended up joining the Jhansi Regiment in Azad Hind Fauz after the marriage. It was the regiment that the British always accused them of spying on British Govt.

It was the time when Neera's husband Jairanjan Das was given the responsibility to spy on Netaji Subhash..

Important paper from Google on large batch optimization. They do impressively careful experiments measuring # iterations needed to achieve target validation error at various batch sizes. The main "surprise" is the lack of surprises. [thread]

https://t.co/7QIx5CFdfJ

The paper is a good example of lots of elements of good experimental design. They validate their metric by showing lots of variants give consistent results. They tune hyperparamters separately for each condition, check that optimum isn't at the endpoints, and measure sensitivity.

They have separate experiments where the hold fixed # iterations and # epochs, which (as they explain) measure very different things. They avoid confounds, such as batch norm's artificial dependence between batch size and regularization strength.

When the experiments are done carefully enough, the results are remarkably consistent between different datasets and architectures. Qualitatively, MNIST behaves just like ImageNet.

Importantly, they don't find any evidence for a "sharp/flat optima" effect whereby better optimization leads to worse final results. They have a good discussion of experimental artifacts/confounds in past papers where such effects were reported.