Anomaly Detection: Algorithms, Explanations, Applications

Anomaly detection is important for data cleaning, cybersecurity, and robust AI systems. This talk will review recent work in our group on (a) benchmarking existing algorithms, (b) developing a theoretical understanding of their behavior, (c) explaining anomaly "alarms" to a data analyst, and (d) interactively re-ranking candidate anomalies in response to analyst feedback. Then the talk will describe two applications: (a) detecting and diagnosing sensor failures in weather networks and (b) open category detection in supervised learning.
See more at www.microsoft....

[@colouredtime]

Thank you for sharing, very interesting lecture. Slides should be shown more often; it is difficult to follow the discussion without seeing the slides.

[@kadenkylo4448]

sorry to be so offtopic but does anyone know a trick to get back into an instagram account??
I was dumb forgot the account password. I would appreciate any help you can give me.

[@emorytony7486]

@Kaden Kylo instablaster ;)

[@kadenkylo4448]

@Emory Tony I really appreciate your reply. I found the site on google and I'm in the hacking process atm.
I see it takes a while so I will get back to you later with my results.

[@kadenkylo4448]

@Emory Tony it worked and I actually got access to my account again. Im so happy!
Thanks so much you saved my ass !

[@emorytony7486]

@Kaden Kylo You are welcome =)

[@MrChirros]

Very difficult to follow without slides. You need to show them more frequently and longer.

[@malharjajoo7393]

Applications start at 1:02:08.
Especially 1:06:50 - Open category classification

[@BollywoodBangg]

Although speakers are very talented and SME, Somehow, The lecture does mediocre job in transform it into great tutorial on subject. it fails on basic editing..(like editor chose to show professor's face lot more than the "slide content" he is showing). Add to misery, ugly behaviour from youtube when you pause video for slides, it shows lot of junk in 50% of screen content killing slide material. Needless to say, video fails to do its job and could have been better...lot better.

[@geoffreyanderson4719]

This is a nice intro to AD theory. Viewers need to see the slideshow more than the instructor himself though. Videographer please take note.

[@nameunknown007]

You should change your cameraman, he has no clue what is happening or how to record a technical discussion. What a huge loss. If he doesn't speak English, maybe he should just go with half screen or bottom screen to show slides. No offense sir, but I don't know why the cameraman thinks the audience is interested in seeing the speaker's face expressions!

[@usamarahman6178]

the lecture was really good but either show the slides, or atleast share the slides

[@wei-chunlee7140]

These talks are great, but showing slides are usually more useful...

[@cheukkinpoon4428]

this video edit is very bad. For most of the time the speaker is talking about the slide, the video is showing the speaker's face instead of the slide like his facial expression is more important than the content he is talking about...

[@DF-ip2yr]

Interesting topic but terrible, terrible recording. What the hell, why you show the speaker more than slides? This is the most terrible way to record a technical lecture. He's talking and we need to imagine what he means.

[@alwaaffa]

You can help me with a master’s thesis for my software part (coding) in Python?

[@sau002]

Where are the slides? They are very important. Becomes hard to follow the speaker.

[@АндрейЗахаров-б3ф]

I've found the slides via google in pdf: web.engr.oregonstate.edu/~tgd/talks/dietterich-anomaly-detection-workshop-2016.pdf

[@3osa16]

@@АндрейЗахаров-б3ф good effort but they aren't the same

[@mallibus]

@@3osa16 yes they are different but they are good enough

[@loicverbeke216]

Is there a link where I can find the slides of the presentation ?

[@benjaminstk]

good presentation!

[@EduardoMosca-t4p]

Where can one find the paper on this(if there is one)?

[@shawn-jung]

Thank you for sharing the session video!

[@_N0_0ne]

Thank you kindly

[@diegocerda6185]

Muy bueno, muchas gracias.

[@anryxas1]

Very interesting area of ML. Thank you for the video

[@trucllii3537]

Loda 51:20

[@thecrabempire5767]

you just have to have a bolt

[@trustmebaz]

Thank you for sharing

[@qrubmeeaz]

Great talk. Thank you.

[@LanDiEvil]

With the knowledge gifted upon me by this video, my algorithmic program to predict the Suns rotation around the Earth may finally come to fruition. Then as darkness falls and everyone stumbles blindly into the night, my path shall be illuminated by a flashlight that I knew to activate moments before the sun sets. The superhuman ability of preparatory night vision will strike fear into the hearts of my enemies and before long, the nations of Earth shall crown me Supreme Global Champion King of Darkness Infinity and feed me grapes as I command an army of robots against the civilian population from a solid gold lounge chair. In a Sky fortress with laser turrets and a front door that looks like a giant skull. And each blade of grass on my front lawn will be hand carved from enriched plutonium, by glowing monks. And to counter the inevitable collapse of civilisation, I shall appoint AI to all levels of government, effectively saving the world and securing my position as 'Ultra-Global Supreme Champion King Hyper-Infinity-Master" long after my death.

[@SidTrip]

dafuq did I just read

[@shantanudave4692]

Hi, myself Dark AI, where can I apply for a job with you :p

[@rvmods]

Thanks for share.