addressalign-toparrow-leftarrow-rightbackbellblockcalendarcameraccwcheckchevron-downchevron-leftchevron-rightchevron-small-downchevron-small-leftchevron-small-rightchevron-small-upchevron-upcircle-with-checkcircle-with-crosscircle-with-pluscrossdots-three-verticaleditemptyheartexporteye-with-lineeyefacebookfolderfullheartglobegmailgooglegroupsimageimagesinstagramlinklocation-pinm-swarmSearchmailmessagesminusmoremuplabelShape 3 + Rectangle 1outlookpersonJoin Group on CardStartprice-ribbonShapeShapeShapeImported LayersImported LayersImported Layersshieldstartickettrashtriangle-downtriangle-uptwitteruserwarningyahoo

November 2016 Speaker Series

Space and food for this month's event has graciously been provided by Tonic Design

Agenda: 
6:00: Meet and greet 
6:30: Sebastian Raschka 
7:00: Lightning talks (Interested in speaking?)

Title: 
Machine Learning and Performance Evaluation — Overcoming the Selection Bias 

Abstract: 
Every day in scientific research and business applications, we rely on statistics and machine learning as support tools for predictive modeling. To satisfy our desire to model uncertainty, to predict trends, and to predict patterns that may occur in the future, we developed a vast library of tools for decision making. In other words, we learned to take advantage of computers to replicate the real world, making intuitive decisions more quantitative, labeling unlabeled data, predicting trends, and ultimately trying to predict the future. Now, whether we are applying predictive modeling techniques to our research or business problems, we want to make "good" predictions!

In the presence of modern machine learning libraries, choosing a machine learning algorithm to fit a model to our training data has never been that simple. However, making sure that our model generalizes well to unseen data is still up to us—the machine learning practitioners and researchers. In this talk, we will discuss the two most important components of estimating generalization performance: bias and variance. We will discuss how we can make the best use of our data at hand—via proper (re)sampling—and how to pick appropriate performance metrics. Then, we will compare various techniques for algorithm selection and model selection to find the right tool and approach for our task at hand. In the context of the "bias-variance trade-off," we will go over potential weaknesses in common modeling techniques, and we will learn how to take uncertainty into account to build predictive model performs well on unseen data.

Bio: 
Sebastian Raschka is the author of the bestselling book Python Machine Learning. As a Ph.D. candidate at Michigan State University, he is developing new computational methods in the field of computational biology. Sebastian has many years of experience with coding in Python and has given several seminars on the practical applications of data science and machine learning. Sebastian loves to write and talk about data science, machine learning, and Python, and he is really motivated to help people developing data-driven solutions without necessarily requiring a machine learning background.

Sebastian is also actively contributing to open source projects, and methods that he implemented are now successfully used in machine learning competitions such as Kaggle. In his free time, Sebastian is also working on models for sports predictions, and if he is not sitting in front of a computer, he enjoys playing sports in his spare time. 



Join or login to comment.

  • Don G.

    Great talk. Thanks to all the organizers. FYI, after the talk I asked Sebastian about the state of the art of the interpretability of models and he pointed me to LIME (see https://www.oreilly.com/learning/introduction-to-local-interpretable-model-agnostic-explanations-lime). He also pointed me to a free Coursera course entitled, "Neural Networks for Machine Learning" by Geoff Hinton. Very cool!

    1 · December 2

  • Keith A.

    Link to my project site.
    http://unicorninvesting.us/

    Link to the FCNN4R presentation that I went through in the lightning talk. Feedback is appreciated.
    http://unicorninvesting.us/content/fast-compressed-neural-network-r-evaluation Feel free to message me.

    2 · November 30

  • Michael B.

    Slides for yesterdays talk are available at https://speakerdeck.com/rasbt/machine-learning-and-performance-evaluation-at-dataphilly-2016 and video is available at https://youtu.be/JlctsNhlNaI

    Thanks again to Sebastian and Keith for speaking and to Tonic for hosting! Don't forget to check out our upcoming events!

    5 · December 1

  • Michael B.

    Hi Folks. Google Hangouts Live is having an outage so unfortunately we won't be live streaming tonight.

    November 30

    • sashi

      I would have loved this... Meh.. I would have downloaded and watched this again and again hmm

      December 1

    • Michael B.

      The speaker attempted to record things on his end so we might be able to upload his talk at some point in the future. I'll post here if it becomes available.

      1 · December 1

  • Anders B.

    Will there be a presentation recording or a powerpoint for download? I could not make it to this one.

    December 1

  • Les

    Informative presentation

    December 1

  • Georgi S.

    The speaker was very well prepared. The topic was excellent for this meetup. It was appropriate for people with different knowledge levels. The lighting talk was a nice addition.

    1 · November 30

  • Ben L.

    Thanks to Tonic Design, Sebastian, Keith, and organizers for a well put together and insightful meetup!

    1 · November 30

  • lila

    Live link?

    November 30

  • Michael B.

    For anyone planning on attending tonight just a reminder that to get to the 3rd floor you'll need to use the secret back elevator (closest to 4th).

    November 30

    • Georgi S.

      If you see 2 elevators that is the wrong place. Go down the hall way.

      November 30

  • Arjun K.

    Hi Michael, Randy, are there open slots for the lightning talks?

    November 22

    • Michael B.

      I believe so. If you'd like to give a 5-15 minute lightning talk there should be time available at the end of the event.

      November 30

  • Ted S.

    I am very interested in attending this but need to make sure I'll be able to get into the city at that time before RSVPing. As a very early-on novice, is there any specialized knowledge I need to have before attending?

    November 16

    • Ted S.

      Thank you, I appreciate your response. I'm still too new for all that, still learning to code and brushing up on my stats.

      November 28

    • Randy O.

      I think you'll still learn a good bit from the talk regardless. Model selection is important to know about, even from a conceptual level.

      1 · November 29

  • Matthew H.

    It seems the address has been cut off a bit...

    Tonic Design
    441 N 5th St, Suite 301, PA

    It should have Philadelphia, PA 19123 in addition.

    2 · November 13

Our Sponsors

People in this
Meetup are also in:

Sign up

Meetup members, Log in

By clicking "Sign up" or "Sign up using Facebook", you confirm that you accept our Terms of Service & Privacy Policy