BC DATA SCIENCE

Saved by @Jacobtldr

Greg Yang
@TheGregYang 5 years, 9 months ago 903 views

Save to PDF Share See On Twitter

1/ A ∞-wide NN of any architecture is a Gaussian process (GP) at init. The NN in fact evolves linearly in function space under SGD, so is a GP at any time during training. https://t.co/v1b6kndqCk With Tensor Programs, we can calculate this time-evolving GP w/o trainin any NN

2/ In this gif, narrow relu networks have high probability of initializing near the 0 function (because of relu) and getting stuck. This causes the function distribution to become multi-modal over time. However, for wide relu networks this is not an issue.

3/ This time-evolving GP depends on two kernels: the kernel describing the GP at init, and the kernel describing the linear evolution of this GP. The former is the NNGP kernel, and the latter is the Neural Tangent Kernel (NTK).

4/ Once we have these two kernels, we can derive the GP mean and covariance at any time t via straightforward linear algebra.

5/ So it remains to calculate the NNGP kernel and NT kernel for any given architecture. The first is described in https://t.co/cFWfNC5ALC and in this thread
https://t.co/6RO7VZDQNZ

6/ The NTK for any architecture is calculated in https://t.co/v1b6kndqCk and in this thread
https://t.co/OOoOMdPOsR

More from Data science

Research Engineering a...
@turinghut23

✨✨ BIG NEWS: We are hiring!! ✨✨
Amazing Research Software Engineer / Research Data Scientist positions within the @turinghut23 group at the @turinginst, at Standard (permanent) and Junior levels 🤩

👇 Here below a thread on who we are and what we

We are a highly diverse and interdisciplinary group of around 30 research software engineers and data scientists 😎💻 👉 https://t.co/KcSVMb89yx #RSEng

We value expertise across many domains - members of our group have backgrounds in psychology, mathematics, digital humanities, biology, astrophysics and many other areas 🧬📖🧪📈🗺️⚕️🪐
https://t.co/zjoQDGxKHq
/ @DavidBeavan @LivingwMachines

In our everyday job we turn cutting edge research into professionally usable software tools. Check out @evelgab's #LambdaDays 👩‍💻 presentation for some examples:

We create software packages to analyse data in a readable, reliable and reproducible fashion and contribute to the #opensource community, as @drsarahlgibson highlights in her contributions to @mybinderteam and @turingway: https://t.co/pRqXtFpYXq #ResearchSoftwareHour

Pat Schloss
@PatSchloss

Wellll... A few weeks back I started working on a tutorial for our lab's Code Club on how to make shitty graphs. It was too dispiriting and I balked. A twitter workshop with figures and code:

When are you doing pie charts?
— #BlackLivesMatter (@surt_lab) October 13, 2020

Here's the code to generate the data frame. You can get the "raw" data from https://t.co/jcTE5t0uBT

Obligatory stacked bar chart that hides any sense of variation in the data

Obligatory stacked bar chart that shows all the things and yet shows absolutely nothing at the same time

STACKED Donut plot. Who doesn't want a donut? Who wouldn't want a stack of them!?! This took forever to render and looked worse than it should because coord_polar doesn't do scales="free_x".

Advertisement

Goku Mohandas
@GokuMohandas

🔥 Putting ML in Production! We're going to publicly develop @madewithml's first ML service. Here is the broad curriculum:

- 📦 Product
- 🔢 Data
- 🤖 Modeling
- 📝 Scripting
- 🛠 API
- 🚀 Production

More details (lessons, task, etc.) here: https://t.co/xmMm9XGK9j

Thread 👇

Questions that this thread will answer:

- What is it?
- Who is this course for?
- What is the format?
- What makes this course unique?
- Why constrain to open source tools?
- What are my qualifications?
- Why is this free?
- What are the

What is it?

Putting ML in Production: a guide and code-driven case study on MLOps. We will be developing and deploying Made With ML's first ML service, from Product → ML → Production, with open source tools.

This ML service will act as a foundation for all future ML features and subsequent iterations. The first feature is tagifai - multilabel classification of tags for a project. We'll discuss the need and utility of this feature in the first lesson.

Who is this course for?

- ML developers looking to become end-to-end ML developers.
- Software engineers looking to learn how to responsibly deploy and monitor ML systems.
- Product managers who want to have a comprehensive understanding of the different stages of ML dev.

Jeremy Howard
@jeremyphoward

An amazing new project from @bearpelican was just released: https://t.co/DBov6sZTVS . A beautiful design; you can auto-generate a melody from chords, chords from a melody, and more.

It's technically brilliant, combining BERT, seq2seq, and Transformer XL
https://t.co/jF3mO5aXiu

It's also a wonderful example of leveraging and customizing the fastai framework in a deep & thoughtful way.

Here's the full set of blog posts diving in to this

Ryan J. Gallagher
@ryanjgallag

Tired of word clouds? Want to do better sentiment analysis? Not sure how to look at the words underneath your measures?

Our long overdue paper on generalized word shift graphs is finally here!
https://t.co/lIBXvbMJWX
https://t.co/vSL1REYT8V

So what are they?

1/n

If we have two texts, there are many ways we can compare them. Weighted averages are a particularly useful measure because they're flexible and interpretable

Proportions, Shannon entropy, the KLD, the JSD, and dictionary methods can all be written as weighted averages

2/n

But weighted avgs are also slippery. When we try to compress complex phenomena like happiness, surprise, divergence, or diversity into a single number, it can be unclear what we're measuring

If the measure goes up, what does that mean? Why did it do that? Can we trust it?

3/n

Very often, that's the end of the line and we're left with an uneasy feeling in the pit of our stomach that our weighted avg is actually picking up a data artifact or some other unintended peculiarity

Word shift graphs help us address those concerns

4/n

First, word shifts look under the hood of weighted averages to see what's going on

All weighted averages are a sum of contributions from individual words. We can pull out those words, and rank which ones contribute the most to the difference between two texts

5/n

You May Also Like

$\u0b85\u0bb8\u0bcd\u0ba4\u0bbf\u0ba9\u0bbe\u0baa\u0bc1\u0bb0\u0ba4\u0bcd\u0ba4\u0bc1 \u0b85\u0bb0\u0b9a\u0bbf\U0001f478$

அஸ்தினாபுரத்து அரசி👸...
@SivaRoobini555

#தினம்_ஒரு_திருவாசகம்
தொல்லை இரும்பிறவிச் சூழும் தளை நீக்கி
அல்லல் அறுத்து ஆனந்தம் ஆக்கியதே – எல்லை
மருவா நெறியளிக்கும் வாதவூர் எங்கோன்
திருவாசகம் என்னும் தேன்

பொருள்:
1.எப்போது ஆரம்பித்தது என அறியப்படமுடியாத தொலை காலமாக (தொல்லை)

2. இருந்து வரும் (இரும்)

3.பிறவிப் பயணத்திலே ஆழ்த்துகின்ற (பிறவி சூழும்)

4.அறியாமையாகிய இடரை (தளை)

5.அகற்றி (நீக்கி),

6.அதன் விளைவால் சுகதுக்கமெனும் துயரங்கள் விலக (அல்லல் அறுத்து),

7.முழுநிறைவாய்த் தன்னுளே இறைவனை உணர்த்துவதே (ஆனந்த மாக்கியதே),

8.பிறந்து இறக்கும் காலவெளிகளில் (எல்லை)

9.பிணைக்காமல் (மருவா)

10.காக்கும் மெய்யறிவினைத் தருகின்ற (நெறியளிக்கும்),

11.என் தலைவனான மாணிக்க வாசகரின் (வாதவூரெங்கோன்)

12.திருவாசகம் எனும் தேன் (திருவா சகமென்னுந் தேன்)

முதல்வரி: பிறவி என்பது முன்வினை விதையால் முளைப்பதோர் பெருமரம். அந்த ‘முன்வினை’ எங்கு ஆரம்பித்தது எனச் சொல்ல இயலாது. ஆனால் ‘அறியாமை’ ஒன்றே ஆசைக்கும்,, அச்சத்துக்கும் காரணம் என்பதால், அவையே வினைகளை விளைவிப்பன என்பதால், தொடர்ந்து வரும் பிறவிகளுக்கு, ‘அறியாமையே’ காரணம்

அறியாமைக்கு ஆரம்பம் கிடையாது. நமக்கு ஒரு பொருளைப் பற்றிய அறிவு எப்போதிருந்து இல்லை? அதைச் சொல்ல முடியாது. அதனாலேதான் முதலடியில், ஆரம்பமில்லாத அஞ்ஞானத்தை பிறவிகளுக்குக் காரணமாகச் சொல்லியது. ஆனால் அறியாமை, அறிவின் எழுச்சியால், அப்போதே முடிந்து விடும்.

Angad Singh
@Singh7575

𝙎𝙝𝙖𝙧𝙞𝙣𝙜 𝙢𝙮 𝙬𝙞𝙨𝙙𝙤𝙢 𝑾𝒐𝒏'𝒕 𝒃𝒆 𝒔𝒖𝒓𝒑𝒓𝒊𝒔𝒆𝒅 𝒊𝒇 𝒕𝒐𝒎𝒐𝒓𝒓𝒐𝒘 𝒖 𝒓𝒆𝒂𝒅 𝒕𝒉𝒆 𝒔𝒂𝒎𝒆 𝒔𝒕𝒖𝒇𝒇 𝒊𝒏 50𝒌 𝒘𝒐𝒓𝒌𝒔𝒉𝒐𝒑 𝒐𝒓 𝒔𝒐𝒎𝒆𝒐𝒏𝒆 𝒎𝒂𝒏𝒂𝒈𝒊𝒏𝒈 𝒚𝒐𝒖𝒓 𝒎𝒐𝒏𝒆𝒚 𝒘𝒊𝒕𝒉 𝒔𝒂𝒎𝒆 𝒍𝒐𝒈𝒊𝒄
Simple and effective way 2 make Money

Idea 1:- Use pivot level like 14800 in case of nifty and sell 14800straddle monthly expiry (365+335) exit if nifty closes on daily basis below S1 or above R1

After closing below S1 if it closes above S1 next day or any day enter the same position again vice versa for R1

Idea2:- Use R1 and S1 corresponding strikes multiple
Incase of R1 15337 take 15300ce
N in case of S1 14221 use 14200pe
Sell both and hold till expiry or exit if nifty closes below S1 or above R1 around closing
If the same bounces above S1 and falls below R1 re-enfer same strikes

Use same criteria for nifty, usdinr and banknifty

(This is must)Use this margin rule for 1lot banknifty pair keep 4Lax margin
For nifty one lot keep 3Lax
For usdinr 100lots keep 4Lax

I bet you if you do this on consistent basis your ROI will be more than 70% on yearly basis.

Couldn't explain easier than this

Criticisms are most welcomed.

Advertisement

$Lu_lu\U0001f498$

Lu_lu💘
@luciame16

Some thoughts on episode 53 of #Hercai

Love,that is not real,has no passion. Its fire turns into embers, it doesn’t not smell. The day, the night, the moon&year burn without sun, light, comfort&sleep.
Ömer Hayyam

#Reyyan #Miran #ReyMir #EbruŞahin #AkınAkınözü
#HercaiThoughts

1
PRAISE
The praise in this episode goes foremost to Ebru because she portrayed Reyyan so realistically like a woman who truly lost her unborn child. I was in awe and pain watching her along the episode, all her breakdowns and tears were so real.
#Hercai

2
Of course also Akın took his part and did his part masterfully, but he rightfully was the one who stood aside in this specific episode that needed to be about Reyyan’s pain and thus concentrated on Ebru.
#Hercai

3
And last but not least, if I say thank you to something that B gave us this year, then it’s to have Ali as our director. Also in this episode we saw the artistic eye behind every shot and I loved it.
#Hercai

4
OVERALL COMMENT
I don’t know who wrote ep. 52 but ep. 53 was going strong once more on symbols: roses, stones, fire and trees were incorporated and especially the ReyMir dialogues were written very well. I truly loved every ReyMir scene.
So let's get into it.
#Hercai

WORLD OF SANATAN DHARM...
@world_sanatan

THE LAST STAND OF PAVAN KHIND - THE 300 BATTLE OF INDIA

A long🧵that u must read

In desperate times, great men come forth. Those men who are ready to sacrifice all of their lives work for a single cause.

Baji Prabhu Deshpande is one such man.

1/25

Battle was between the Maratha Warrior Baji Prabhu Deshpande and Sidi Masoud of Adil Shahi Sultanate.

In 1660, the Maratha King Shri Chhatrapati Shivaji was trapped in the fort of Panhala, under siege and vastly outnumbered by an Adilshahis army led by Sidi Masoud.

2/25

And while everyone is thinking that finally the Maratha king will be subdued, Shivaji planned to escape to the fort of Vishalgadh. It was administered by a Maratha chieftain named Range Narayan Orpe under Shivaji.

3/25

For months the siege warfare continued as Shivaji made sure that the food source of the Besieging army drops down, due to which they will initiate a general assault on the castle.

4/25

Shivaji, Baji Prabhu, and around 600 of their best troops, hardened mountaineers of the Maval region, would dash through the Adilshahis force at night.

A man named Shiva Kashid, who resembled Shivaji in appearance, had volunteered to dress like the king and get captured.

5/25

$\u092a\u094d\u0930\u0924\u093f\u091c\u094d\u091e$

प्रतिज्ञ...
@RamaInExile

#twitterArts
Arts are best form to Express thoughts & devotion. In All civilization Arts played very important role to express sprituality. Tibbeten Tangkhas are prime example .

List of My Paintings & Sketches

प्रलय पयोधि-जले धृतवान् असि वेदम्
विहित वहित्र-चरित्रम् अखेदम्
केशव धृत-मीन-शरीर, जय जगदीश हरे!!

छलयसि विक्रमणे बलिम् अद्भुत-वामन
पद-नख-नीर-जनित-जन-पावन
केशव धृत-वामन रूप जय जगदीश हरे

नंदसि यज्ञ- विधेर् अहः श्रुति जातम्
सदय-हृदय-दर्शित-पशु-घातम्
केशव धृत-बुद्ध-शरीर जय जगदीश हरे

क्षत्रिय-रुधिर-मये जगद् -अपगत-पापम्
स्नपयसि पयसि शमित-भव-तापम्
केशव धृत-भृगुपति रूप जय जगदीश हरे