Data Platform for the Software Development Life Cycle. #EngineeringEffectiveness #MLonCode⠀| GitHub: https://xitter.cf/t.co/4XuUmDJURU⠀| website: https://xitter.cf/t.co/Gvetg0CBGy
Joined November 2015
- Tweets1.8K
- Following1.3K
- Followers3.1K
- Likes1.1K
source{d} retweeted
👏 Not only for architects but for a whole organization This is what I am used to call #CodeAsData from a few years.
Hope to be able to spend time on it very soon! #MLOnCode @eisokant @natfriedman @sourcedtech @Semmle
You’re unable to view this Post because this account owner limits who can view their Posts. Learn more
source{d} retweeted
“Companies focused on modernizing their software stacks, getting rid of technical debt, and embracing #OpenSource software development and #DevOps best practices will find source{d} extremely valuable.” @mazoff, @Ovum buff.ly/2BSRVGS
source{d} retweeted
.@omojumiller, @HamelHusain and @jeremylewi have talked about #MLonCode datasets and applications on #TFWorld. @sourcedtech should participate in the CodeSearchNet competition! The issue labeling bot looks awesome.
source{d} retweeted
How “MLonCode” Could Disrupt Software Development ow.ly/pHCj30pIBZx @sourcedtech #machinelearning #softwaredevelopment
source{d} retweeted
Great presentation of Apple's Overton by Hugo Mougard, @sourcedtech's ML team member, on Nantes Machine Learning Meetup: github.com/nantes-machine-le… I love how they "patch" their deep learning models and combine weak labelers with Snorkel.
source{d} retweeted
I have put some real effort into this.
Multi-GPU #DeepLearning at @sourcedtech buff.ly/2AXsNxW by @vadimlearning
🚀 Learn how we solved several problems in order to train #NeuralNetworks with @TensorFlow 2.0 on several local #GPUs in our #ML cluster 🚀
source{d} retweeted
Here's an abridged version of my ASE'19 paper: tinyurl.com/davisetal-ase19
Still too long? TL;DR:
(1) As long as you look at enough software projects, you can extract regexes using static analysis or program instrumentation.
(2) Regexes are similar across programming languages.
source{d} retweeted
"How Bugs Are Born: A Model to Identify How Bugs Are Introduced in Software Components" by @Gerope90 @gregoriorobles @aserebrenik Daniel M German @azaidman @jgbarah accepted for @emsejournal. Preprint available here: azaidman.github.io/publicati…
🇨🇦🇪🇸🇳🇱
source{d} retweeted
The engineering demos concluded my first day at @sourcedtech 🤩
source{d} retweeted
🎉 The wait is over! TensorFlow 2.0 is finally here.
Driven by community feedback, this release provides a complete set of tools for developers, enterprises, and researchers to easily build ML applications.
Read the blog ↓ goo.gle/TF2rel
source{d} retweeted
"I-MAD: A Novel Interpretable Malware Detector Using Hierarchical Transformer" by Miles L., Benjamin F., Philippe C., Steven D.
arxiv.org/abs/1909.06865
Stacked custom Transformers pre-trained on Masked Assembly LM for malware classification & Assembly Fn Clone Detection tasks
source{d} retweeted
🤗Transformers 2.0💥
State-of-the-art NLP in TensorFlow 2.0/PyTorch
8 architectures
33 trained models
102 lang.
Seamlessly pick the right framework for training, eval, deploy
Train on TPU ⏩ finetune/test in PyTorch ⏩ serve w. TFX
🍒Keras magic: train SOTA model in 10 lines👇
source{d} retweeted
Looking to write more professional looking code? Join Kaggle data scientist @rctatman tomorrow as she shares some of her top tips and tricks for writing reusable code. ow.ly/hsrM50wpRPt
You can check out her notebook ahead of time here: ow.ly/LBVi50wpRP9
source{d} retweeted
Honored to discuss my and @eiso's paper about #MLonCode from the good old times at @sourcedtech in TU Delft together with @gousiosg's and @mauricioaniche's students
source{d} retweeted
Yesterday we had an awesome time at #madScalability talking about @kubernetesio, microservices and github.com/mcuadros/ascode.
Thanks to @autentia (videos are coming), @seedtag (@Laura_Morillo & @paulgoldbaum) and @sourcedtech (@mcuadros_) 👏🏽
Nov is coming! meetup.com/madscalability/
source{d} retweeted
Great to see the reference in TechCrunch how @cloudfoundry is taking advantage of @sourcedtech!
techcrunch.com/2019/09/12/to…
source{d} retweeted
"End-to-end Deep Learning of Optimization Heuristics" by @iamchriscummins, Zheng Wang, Hugh Leather.
- 📃homepages.inf.ed.ac.uk/hleat…
- 💻github.com/ChrisCummins/pape…
Join @sourcedtech reading club on 2019.09.20 at 4PM CET to discuss it with one of the authors github.com/src-d/reading-clu…
source{d} retweeted
Check out our work "SEQUENCER: Sequence-to-Sequence Learning for End-to-End Program Repair" accepted at TSE by Zimin Chen, Steve James Kommrusch, @tufanomichele, Louis-Noël Pouchet, Denys Poshyvanyk, @martinmonperrus.
Paper: arxiv.org/abs/1901.01808
source{d} retweeted
Slides of the talk "SE applications with ML aspects" available at tinyurl.com/aise19
At AISE workshop @timofeybryksin of @jetbrains talking about data collection for #MLonCode including the PGA dataset github.com/src-d/datasets/tr… provided by @sourcedtech
source{d} retweeted
At AISE workshop @timofeybryksin of @jetbrains talking about data collection for #MLonCode including the PGA dataset github.com/src-d/datasets/tr… provided by @sourcedtech