Tweetovi

Blokirali ste korisnika/cu @MohdShukriHasa1

Jeste li sigurni da želite vidjeti te tweetove? Time nećete deblokirati korisnika/cu @MohdShukriHasa1

  1. 3. velj

    This 1-hour talk from Salesforce's Director of Data Science is packed with lots of practical tips and advice on what it means to do "Agile Data Science".

    Poništi
  2. 2. velj

    Today I learned a new method to explain the predictions from black box models called counterfactuals. Given an input, x, and its predicted output y, this method perturbs x to get x' just enough so that y' is different from y. You can then compare x with…

    Poništi
  3. 31. sij

    Are you a CTO/CIO/SVP/Director looking to build a world-class data science team? Here are 5 tips to help you get started: 1. Decide what kind of data scientists you want. Do you want someone to help infuse AI into your products, optimize your operations…

    Poništi
  4. 29. sij

    I really enjoyed reading this blog post from Chris Rackauckas on a method to embed prior knowledge into a neural network to make it less data-hungry. It addresses a very important and challenging problem in a way accessible to peop…

    Poništi
  5. 27. sij

    Checkout this report on DataOps (an Agile take on data engineering) from Oreilly. Very insightful if you are currently building (or thinking about it) a team to execute your organization's data strategy. Surprisingly, the authors…

    Poništi
  6. 20. sij

    I made it!!! I'm now a Certified Professional! If you ❤ Graph DBs, I challenge you to take the exam too:

    Poništi
  7. 19. sij

    There's a grant meant to support open source software critical to biomedical research. Past recipients include Bioconductor, numpy, and matplotlib to name a few. Notably absent is anything from the Julia ecosystem. Why? I think it…

    Poništi
  8. 18. sij

    Just found this well put together and detailed tutorial on using BERT for document classification. Very useful if you need a crash course on using BERT for your own use cases.

    Poništi
  9. 18. sij

    A tip for data scientists who write code for production: If you find yourself relying on extra if...else statements as you add more models into your system, you are likely setting yourself up for a maintenance nightmare. Check out…

    Poništi
  10. 16. sij

    AI is the new electricity (Andrew Ng, 2016) and everyone is well aware of the transformative effects electricity has had across all industries. So, it is only natural to expect AI-based companies to be selling at rich "valuations" given their potential to…

    Poništi
  11. 13. sij

    How can companies get their developers to embrace AI? Take a page from Morningstar and organize a corporate-wide AI competition, of course! Just came across this news article that Morningstar organized an internal global AWS DeepRacer competition (link i…

    Poništi
  12. 12. sij

    I've met with a few large, well-established companies looking to hire their first data scientist. Although the key decision-makers were enthusiastic about adopting AI, none could articulate what their AI strategy was. This got me thinking: How should matu…

    Poništi
  13. 8. sij

    If you are a data scientist, you have obviously committed to lifelong learning. But do you possess the skills to learn efficiently? Being able to assimilate new knowledge fast is a valuable skill given that many of us struggle jug…

    Poništi
  14. 7. sij

    What's the next step after you discovered a superior model from your machine learning experiments? Package your code into a docker container and pass it to the folks at production for deployment (or maybe even deploy it yourself)? Seems like this is the m…

    Poništi
  15. 2. sij

    It's the new year so this seems like an apt podcast to share. Learn what the latest research says about forming long-lasting good habits.

    Poništi
  16. 1. sij

    Folks in software engineering have well-defined steps to test the software they write. How can these practices be transferred to testing a machine learning system? I came across this talk (link in the description), that breaks an ML system into 4 stages:…

    Poništi
  17. 30. pro 2019.

    Building a production-grade recommender system is hard. There are many aspects you need to get right: dataset split, objective metrics and model evaluation just to name a few. This is just the algorithm design part. Operationalizin…

    Poništi
  18. 30. pro 2019.

    Allen Downey's BayesMadeSimple repo contains many fun exercises to get started with Bayesian analysis. I've translated the solutions to Gen, a probabilistic programming framework written in Julia. Just wanted to raise more awarenes…

    Poništi
  19. 27. pro 2019.

    Newcomers to Spark will learn early on that Spark is lazy. Not keeping this in mind at all times may result in incorrect analyses in subtle ways. Check out this blog post to see how lazy evaluation can mess up your analyses on rand…

    Poništi
  20. 25. pro 2019.

    I really enjoyed watching this webinar from Prof. Downey. It bridges the gap between Grid Approximation and MCMC methods in Bayesian analysis. It's also a practical introduction to working with PyMC3.

    Poništi

Čini se da učitavanje traje već neko vrijeme.

Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.

    Možda bi vam se svidjelo i ovo:

    ·