BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Finding Signals in Twitter with ML/NLP at Bloomberg - Minjie Xu - 
 Bloomberg Software Engineer\, Social Media Analytics
DTSTART:20170425T120000Z
DTEND:20170425T130000Z
UID:TALK72011@talks.cam.ac.uk
CONTACT:Jan Samols
DESCRIPTION:Ever wondered how to:\n•	Find out the most talked-about comp
 anies\, topics on Twitter?\n•	Get alerted when something breaks out?\n
 •	Obtain a quick overview of what's going on (e.g. when Donald Trump pos
 ts again)?\n\nBloomberg is all about up-to-date financial data and analysi
 s. In the current social media era\, Twitter has proven itself an indispen
 sable source of information as we frequently see information posted or sha
 red on Twitter being big market movers.\n\nThanks to a proprietary access 
 deal with Twitter\, we can receive every single tweet that matches a set o
 f rules in near real-time\, on top of which we then perform a series of ML
 /NLP analytics\, extracting meaningful signals and filtering out irrelevan
 t noise. \n\nTo this end\, we have built a range of products around Twitte
 r\, including trending companies\, company sentiment\, topic streams\, etc
 .\n\nIn this talk\, Minjie will be covering several topics:\na) A general 
 overview of how Bloomberg processes Twitter data\nb) Highlight our in-hous
 e Twitter topic streams system\nc) Present several interesting Machine Lea
 rning topics therein\nand if time allows\,\nd) Discuss a more recent proje
 ct that involves learning good document embeddings (via variational auto-e
 ncoders) for use in clustering\n\nFood and the opportunity to network afte
 r the talk in FW26.\n\n\n
LOCATION:LT1\, Computer Laboratory
END:VEVENT
END:VCALENDAR