Sigrid Keydanahttp://blog.trivadis.com/b/sigridkeydana/default.aspxen-USTelligent Community 5.6.583.24393 (Build: 5.6.583.24393)Updatehttp://blog.trivadis.com/b/sigridkeydana/archive/2019/04/08/update.aspxMon, 08 Apr 2019 16:45:43 GMT7f420732-9615-472e-9723-d9bd9f35b01c:184855Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=184855http://blog.trivadis.com/b/sigridkeydana/archive/2019/04/08/update.aspx#commentsI’m an applied researcher at RStudio, where I contribute to the r-tensorflow family of packages (e.g., tfprobability). I write about doing deep learning from R on the TensorFlow for R blog. If you’re interested in DL/ML/probability and R,...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2019/04/08/update.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=184855" width="1" height="1">UncategorizedDeep learning, concepts and frameworks: Find your way through the jungle (talk)http://blog.trivadis.com/b/sigridkeydana/archive/2018/02/06/deep-learning-concepts-and-frameworks-find-your-way-through-the-jungle-talk.aspxTue, 06 Feb 2018 14:35:52 GMT7f420732-9615-472e-9723-d9bd9f35b01c:184456Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=184456http://blog.trivadis.com/b/sigridkeydana/archive/2018/02/06/deep-learning-concepts-and-frameworks-find-your-way-through-the-jungle-talk.aspx#commentsToday at OOP in Munich, I had an in-depth talk on deep learning, including applications, basic concepts as well as practical demos with Tensorflow, Keras and PyTorch. As usual, the slides are on RPubs, split up into 2 parts because of the plenty of images...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2018/02/06/deep-learning-concepts-and-frameworks-find-your-way-through-the-jungle-talk.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=184456" width="1" height="1">Machine LearningData Scienceneural networksNatural Language ProcessingPythonDeep LearningtensorflowKerasbackpropagationpytorchcomputer visionPractical Deep Learning (talk)http://blog.trivadis.com/b/sigridkeydana/archive/2017/12/14/practical-deep-learning-talk.aspxThu, 14 Dec 2017 11:08:11 GMT7f420732-9615-472e-9723-d9bd9f35b01c:184406Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=184406http://blog.trivadis.com/b/sigridkeydana/archive/2017/12/14/practical-deep-learning-talk.aspx#commentsYesterday at IT Tage 2017, I had an introductory-level talk on deep learning. After giving an overview of concepts and frameworks, I zoomed in on the task of image classification using Keras, Tensorflow and PyTorch, not aiming for high classification...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2017/12/14/practical-deep-learning-talk.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=184406" width="1" height="1">Machine LearningData Scienceneural networksPythonDeep LearningtensorflowKeraspytorchI’m a developer, why should I care about matrices or calculus? (talk at MLConference 2017)http://blog.trivadis.com/b/sigridkeydana/archive/2017/12/06/i-m-a-developer-why-should-i-care-about-matrices-or-calculus-talk-at-mlconference-2017.aspxWed, 06 Dec 2017 16:55:28 GMT7f420732-9615-472e-9723-d9bd9f35b01c:184394Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=184394http://blog.trivadis.com/b/sigridkeydana/archive/2017/12/06/i-m-a-developer-why-should-i-care-about-matrices-or-calculus-talk-at-mlconference-2017.aspx#commentsYesterday at ML Conference, which took place this year for the first time, I had a talk on cool bits of calculus and linear algebra that are useful and fun to know if you’re writing code for deep learning and/or machine learning. Originally, the...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2017/12/06/i-m-a-developer-why-should-i-care-about-matrices-or-calculus-talk-at-mlconference-2017.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=184394" width="1" height="1">RMachine LearningData Scienceneural networksDeep Learninglinear algebraadversarial learningmatrix factorizationbackpropagationcalculusPlus/minus what? Let’s talk about uncertainty (talk)http://blog.trivadis.com/b/sigridkeydana/archive/2017/11/25/plus-minus-what-let-s-talk-about-uncertainty-talk.aspxSat, 25 Nov 2017 16:35:59 GMT7f420732-9615-472e-9723-d9bd9f35b01c:184380Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=184380http://blog.trivadis.com/b/sigridkeydana/archive/2017/11/25/plus-minus-what-let-s-talk-about-uncertainty-talk.aspx#commentsLast week at DOAG 2017, I had two talks, one about deep learning with DL4J (slides here) and one about how to communicate uncertainty (or rather: how to construct prediction intervals for various methods / in various frameworks ranging from simple linear...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2017/11/25/plus-minus-what-let-s-talk-about-uncertainty-talk.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=184380" width="1" height="1">StatisticsBayesianMachine LearningData Scienceneural networksARIMAconfidence intervalsprediction intervalslinear regressionGaussian processesbayesian statisticsbootstrapDynamic forecasts – with Bayesian linear models and neural networks (talk at Predictive Analytics World Berlin)http://blog.trivadis.com/b/sigridkeydana/archive/2017/11/15/dynamic-forecasts-with-bayesian-linear-models-and-neural-networks-talk-at-predictive-analytics-world-berlin.aspxWed, 15 Nov 2017 20:10:40 GMT7f420732-9615-472e-9723-d9bd9f35b01c:184369Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=184369http://blog.trivadis.com/b/sigridkeydana/archive/2017/11/15/dynamic-forecasts-with-bayesian-linear-models-and-neural-networks-talk-at-predictive-analytics-world-berlin.aspx#commentsI really wish I had the time to write an article about the conference, instead of just posting the slides! Predictive Analytics World was super inspiring, not just in a technical way but also as to the broader picture of today’s data science / AI...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2017/11/15/dynamic-forecasts-with-bayesian-linear-models-and-neural-networks-talk-at-predictive-analytics-world-berlin.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=184369" width="1" height="1">StatisticsRBayesianMachine LearningData Scienceneural networksforecastingDeep LearningLSTMKalman Filtertime seriesDynamic Linear ModelsDeep Learning with Keras – using R (talk)http://blog.trivadis.com/b/sigridkeydana/archive/2017/11/11/deep-learning-with-keras-using-r-talk.aspxSat, 11 Nov 2017 17:36:21 GMT7f420732-9615-472e-9723-d9bd9f35b01c:184347Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=184347http://blog.trivadis.com/b/sigridkeydana/archive/2017/11/11/deep-learning-with-keras-using-r-talk.aspx#commentsThis week in Kassel, [R]Kenntnistage 2017 took place, organised by EODA. It was all about Data Science (with R, mostly, as you could guess): Speakers presented interesting applications in industry, manufacturing, ecology, journalism and other fields,...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2017/11/11/deep-learning-with-keras-using-r-talk.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=184347" width="1" height="1">RMachine LearningData Scienceneural networksDeep LearningTime series shootout: ARIMA vs. LSTM (talk)http://blog.trivadis.com/b/sigridkeydana/archive/2017/10/08/time-series-shootout-arima-vs-lstm-talk.aspxSun, 08 Oct 2017 05:52:24 GMT7f420732-9615-472e-9723-d9bd9f35b01c:184302Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=184302http://blog.trivadis.com/b/sigridkeydana/archive/2017/10/08/time-series-shootout-arima-vs-lstm-talk.aspx#commentsYesterday, the Munich datageeks Data Day took place. It was a totally fun event – great to see how much is going on, data-science-wise, in and around Munich, and how many people are interested in the topic! (By the way, I think that more than half...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2017/10/08/time-series-shootout-arima-vs-lstm-talk.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=184302" width="1" height="1">StatisticsRMachine LearningData Scienceneural networkstimeseriesforecastingDeep Learningrecurrent neural networkARIMALSTMAutomatic Crack Detection – with Deep Learninghttp://blog.trivadis.com/b/sigridkeydana/archive/2017/09/24/automatic-crack-detection-with-deep-learning.aspxSun, 24 Sep 2017 05:35:23 GMT7f420732-9615-472e-9723-d9bd9f35b01c:184267Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=184267http://blog.trivadis.com/b/sigridkeydana/archive/2017/09/24/automatic-crack-detection-with-deep-learning.aspx#commentsOn Friday at DOAG Big Data Days, I presented one possible application of deep learning: using deep learning for automatic crack detection – with some background theory, a Keras model trained from scratch, and the use of VGG16 pretrained on Imagenet...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2017/09/24/automatic-crack-detection-with-deep-learning.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=184267" width="1" height="1">RMachine LearningData Scienceneural networksDeep Learningimage classificationDeep Learning, deeplearning4j and Outlier Detection: Talks at Trivadis Tech Eventhttp://blog.trivadis.com/b/sigridkeydana/archive/2017/09/18/deep-learning-deeplearning4j-and-outlier-detection-talks-at-trivadis-tech-event.aspxMon, 18 Sep 2017 19:20:42 GMT7f420732-9615-472e-9723-d9bd9f35b01c:184263Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=184263http://blog.trivadis.com/b/sigridkeydana/archive/2017/09/18/deep-learning-deeplearning4j-and-outlier-detection-talks-at-trivadis-tech-event.aspx#commentsLast weekend, another edition of Trivadis Tech Event took place. As usual, it was great fun and a great source of inspiration. I had the occasion to talk about deep learning twice: One talk was an intro to DL4J (deeplearning4j), zooming in on a few aspects...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2017/09/18/deep-learning-deeplearning4j-and-outlier-detection-talks-at-trivadis-tech-event.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=184263" width="1" height="1">JavaStatisticsRMachine LearningData Scienceneural networksPythonDeep LearningDL4Jdeeplearning4jvariational autoencoderanomaly detectionHaskell, R, and HaskellR: Combining the best of two worlds (talk at UseR! 2017)http://blog.trivadis.com/b/sigridkeydana/archive/2017/07/07/haskell-r-and-haskellr-combining-the-best-of-two-worlds-talk-at-user-2017.aspxFri, 07 Jul 2017 17:55:40 GMT7f420732-9615-472e-9723-d9bd9f35b01c:184196Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=184196http://blog.trivadis.com/b/sigridkeydana/archive/2017/07/07/haskell-r-and-haskellr-combining-the-best-of-two-worlds-talk-at-user-2017.aspx#commentsEarlier today, I presented at UseR! 2017 about HaskellR: a great piece of software, developed by Tweag I/O, that allows to seemlessly use R from Haskell. It was my first UseR!, it was a great experience, and if I had the time I’d like to write a...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2017/07/07/haskell-r-and-haskellr-combining-the-best-of-two-worlds-talk-at-user-2017.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=184196" width="1" height="1">RMachine LearningData Scienceneural networksDeep LearningFunctional ProgrammingHaskellHaskellRTime series prediction – with deep learninghttp://blog.trivadis.com/b/sigridkeydana/archive/2017/05/26/time-series-prediction-with-deep-learning.aspxFri, 26 May 2017 14:20:32 GMT7f420732-9615-472e-9723-d9bd9f35b01c:184148Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=184148http://blog.trivadis.com/b/sigridkeydana/archive/2017/05/26/time-series-prediction-with-deep-learning.aspx#commentsMore and more often, and in more and more different areas, deep learning is making its appearance in the world around us. Many small and medium businesses, however, will probably still think – Deep Learning, that’s for Google, Facebook &...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2017/05/26/time-series-prediction-with-deep-learning.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=184148" width="1" height="1">StatisticsRMachine LearningData Scienceneural networkstimeseriesforecastingPythonDeep Learningrecurrent neural networkkerasRKerasARIMALSTMDeep Learning in Action (the less mathy version, this time)http://blog.trivadis.com/b/sigridkeydana/archive/2017/03/24/deep-learning-in-action-the-less-mathy-version-this-time.aspxFri, 24 Mar 2017 13:19:08 GMT7f420732-9615-472e-9723-d9bd9f35b01c:184076Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=184076http://blog.trivadis.com/b/sigridkeydana/archive/2017/03/24/deep-learning-in-action-the-less-mathy-version-this-time.aspx#commentsOn Tuesday at Hochschule München, Fakultät für Informatik and Mathematik I again gave a guest lecture on Deep Learning (RPubs, github, pdf). This time, it was more about applications than about matrices, more about general understanding than about architecture...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2017/03/24/deep-learning-in-action-the-less-mathy-version-this-time.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=184076" width="1" height="1">Machine LearningData Scienceneural networksNatural Language ProcessingPythonclassificationDeep Learningperceptronrecurrent neural networkconvolutional neural networkreinforcement learningdeep reinforcement learningR 4 hackershttp://blog.trivadis.com/b/sigridkeydana/archive/2017/03/20/r-4-hackers.aspxMon, 20 Mar 2017 17:00:56 GMT7f420732-9615-472e-9723-d9bd9f35b01c:184066Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=184066http://blog.trivadis.com/b/sigridkeydana/archive/2017/03/20/r-4-hackers.aspx#commentsYesterday at Trivadis Tech Event, I talked about R for Hackers. It was the first session slot on Sunday morning, it was a crazy, nerdy topic, and yet there were, like, 30 people attending! An emphatic thank you to everyone who came! R a crazy, nerdy topic...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2017/03/20/r-4-hackers.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=184066" width="1" height="1">StatisticsRMachine LearningData Sciencelispobject orientationFunctional Programmings3purrrHaskellDeep Learning in Actionhttp://blog.trivadis.com/b/sigridkeydana/archive/2016/12/23/deep-learning-in-action.aspxFri, 23 Dec 2016 10:13:32 GMT7f420732-9615-472e-9723-d9bd9f35b01c:183979Sigrid Keydana1http://blog.trivadis.com/b/sigridkeydana/rsscomments.aspx?WeblogPostID=183979http://blog.trivadis.com/b/sigridkeydana/archive/2016/12/23/deep-learning-in-action.aspx#commentsOn Wednesday at Hochschule München, Fakultät für Informatik and Mathematik I presented about Deep Learning (nbviewer, github, pdf). Mainly concepts (what’s “deep” in Deep Learning, backpropagation, how to optimize …) and architectures...(<a href="http://blog.trivadis.com/b/sigridkeydana/archive/2016/12/23/deep-learning-in-action.aspx">read more</a>)<img src="http://blog.trivadis.com/aggbug.aspx?PostID=183979" width="1" height="1">Machine LearningData Scienceneural networksword2vecNatural Language Processingword vectorsPythonclassificationDeep Learningtensorflowperceptronrecurrent neural networkconvolutional neural network