James Howard

@DrJHoward

Trust PhD Fellow & Cardiology Registrar at | Machine learning in cardiovascular imaging

London, UK
Vrijeme pridruživanja: travanj 2013.

Tweetovi

Blokirali ste korisnika/cu @DrJHoward

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

  1. Prikvačeni tweet
    28. ožu 2019.

    Pleased to say our paper using neural networks to identify pacemaker/ICD/ILRs from chest X-rays has been published in at - you can also play with it at

    Poništi
  2. proslijedio/la je Tweet
    1. velj

    Our meta-analysis of RV pacing for HOCM was one of the most downloaded EHJ-QCCO articles in 2019. ⁩ ⁦ Top downloaded articles 2019:

    Poništi
  3. proslijedio/la je Tweet
    28. sij

    Nineteen fellows from and showcased research projects aimed at tackling global health problems and disease challenges as part of 4i programme. Read more about their exciting work:

    Poništi
  4. proslijedio/la je Tweet
    23. sij

    Artificial intelligence and the cardiologist: what you need to know for 2020 Article by & explaining why AI is relevant to cardiology & demystifying some of the terminology Open access-

    , , i još njih 5
    Poništi
  5. 17. sij

    I've written an example Kaggle kernel showing the dataset in action. A neural network trained in 10 minutes gets test set accuracies over 90% for the exact model present: (95.6% peak accuracy, though that's cherry picking...)

    Prikaži ovu nit
    Poništi
  6. 17. sij

    It's distributed with the CC BY-NC-SA 4.0 license - we only ask that you cite the original publication to use it

    Prikaži ovu nit
    Poništi
  7. 17. sij

    The dataset is the same 1,676 images across 45 device categories used in the index publication, with the same train/test split. At only 9MB it's easily accessible for people interested in giving deep learning a go.

    Prikaži ovu nit
    Poništi
  8. 17. sij

    Pleased to say we've had ethical confirmation that the dataset from our pacemaker identification paper () can be made open. Is it now available on Kaggle here:

    Prikaži ovu nit
    Poništi
  9. proslijedio/la je Tweet
    19. pro 2019.

    Delighted we have been able to publish this beginners guide to EP! Designed to help people get to grips with the array of equipment and jargon used when they walk into the EP lab for the first time! Collaboration with colleagues from and 🇩🇪!

    , , i još njih 4
    Prikaži ovu nit
    Poništi
  10. 16. stu 2019.

    Can't make it to and feel left out of the excitement? Don't worry! Crack open the drinks, get your local cath lab round and play ISCHEMIA Drinking Bingo! 2 hours to go!

    Poništi
  11. proslijedio/la je Tweet
    8. stu 2019.

    Study by Dr. and colleagues in demonstrates the future role of machine learning and in the , with arterial waveform analysis using neural networks allowing for rapid and accurate identification of damping.

    Poništi
  12. proslijedio/la je Tweet
    30. lis 2019.

    Meet Professor Darrel Francis - our new head of section for Cardiovascular Trials & Epidemiology in - find out why he became a scientist & what he sees on the horizon for the section

    Poništi
  13. proslijedio/la je Tweet
    26. lis 2019.

    When Skynet becomes self-aware in 2022, we should all acknowledge its creator.

    Poništi
  14. 26. lis 2019.

    I hope you found this interesting. I'm happy to provide more examples if people want to give me titles. I might try to get the neural network working online like I did for pacemakers () - though the hardware requirements for this are much steeper!

    Prikaži ovu nit
    Poništi
  15. 26. lis 2019.

    What's remarkable is how "scientific" the abstracts feel. Like I said, at no stage have I told it that an abstract should have a background, or that conclusions come at the end, or that meta-analyses should provide effects sizes and literature searches. It's learnt all this.

    Prikaži ovu nit
    Poništi
  16. 26. lis 2019.

    Finally, I thought I'd try something incendiary. It turns out the prevalence of narcotic use amongst adult cardiologists is 71%! And it must be true, they used eigenvectors! Sorry for the slander, my South American colleagues.

    Prikaži ovu nit
    Poništi
  17. 26. lis 2019.

    I thought I'd try something which I have particularly little experience with: cost-effectiveness analysis. While the conclusion might be correct, I'm not sure it adequately conveys the study's findings... I do enjoy how the significance matches the p values (> vs < 0.05).

    Prikaži ovu nit
    Poništi
  18. 26. lis 2019.

    Next I gave it a title for a meta-analysis. Obviously, the title I chose is ludicrous, but I wanted to see what it did. Amazingly, it decided to put a search strategy in the methods section. It also provides relative risks, though the choice of the modified Rankin Scale...

    Prikaži ovu nit
    Poništi
  19. 26. lis 2019.

    First, I tried to give it a title to a made-up randomised controlled trial. It seems a _bit_ unreasonable to compare renal denervation against apixaban for hypertension. Fasciantingly, it volunteered a clinical trial registration data at the end of the abstract.

    Prikaži ovu nit
    Poništi
  20. 26. lis 2019.

    In each of the screenshots below, the top line shows the title I provided the network with, and everything below that is the network's work. It takes around 30 seconds to generate an abstract, though it took over 24 hours of training the network to get it to this level.

    Prikaži ovu nit
    Poništi
  21. 26. lis 2019.

    I've created a monster. I've re-trained 's GPT2 transformer neural network on the Pubmed/MEDLINE database, so that if I give it an article's title, it spits out a abstract for me. I didn't teach it how to structure an abstract, how long it is, or any of the lingo.

    Prikaži ovu nit
    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:

    ·