News @ 2018-08-31 ... paper rush!

Posted on ven. 31 août 2018 in news by Kazuhiro Terao

There was no meeting this week, but here's a flush of news to avoid a clog.


Next meeting Let's plan on September 5th 11:30AM Fermilab/Chicago (US/Central) time. Please send me what awesome you can present. We all want to see your progress!

Recent Neutrino-ML Papers

  • EXO-200 ... using deep neural nets for energy and position estimation of the signal
    • arXiv:1804.09641 or JINST
    • You can find Manuel, one of authors, on our slack to ask questions!
      Figure 13 from EXO-200 paper showing energy reconstruction using deep learning technique making a better energy estimation.
  • MicroBooNE ... using deep neural nets for semantic segmenting shower vs. track particles, validation on lartpc data!
    • arXiv:1808.07269
    • You can find Kazu (me) and Taritree as authors on our slack to ask questions!
      Figure 21 from MicroBooNE paper U-ResNet classifies electromagnetic shower at the pixel level.
  • Minerva ... using deep neural nets to find the interaction vertex + reduction to a model bias using domain advesarial neural nets
    • arXiv:1808.08332
    • You can find Anushree and Gabe as authors on our slack to ask questions!
      Figure 20 (a) from Minerva paper domain adversarial neural netowrk achieves a better accuracy.
  • ANL Bubble Chamber data re-analysis
    • arXiv:1805.00905
    • I have not read this yet! but was introduced in Teppei's neutrino xs newsletter (if you are into nu-xs and not on the list, I recommend)
  • Everyone dance now ... arXiv:1808.07371
    • As Lisa says, "Geeks and nerds can dance!", very relevant to us all. Make sure you train your network before your next date!

Job/Conference/Funding opportunities

As usual, thanks to Gabe Perdue for sharing many opportunities! Checkout deep.learn.physics slack organization for more updates.