Libby Lange

[Libby Lange @LangeLibby]

for anyone interested in Chicago's experience in predictive policing: https://t.co/2RSW7fbFq0 #jacksonbigdata

[Libby Lange @LangeLibby]

The possibility of positive preemptive law enforcement intervention would probably be highly dependent on community experiences of policing in the past. #jacksonbigdata

[Libby Lange @LangeLibby]

I love the idea of accounting for student satisfaction but wonder how we avoid the sunk cost fallacy. By the time you get to the end, you probably want to think the thing you spent $$$ on over 4 years was worthwhile! #jacksonbigdata

[Libby Lange @LangeLibby]

That super complex graphic would have made a good Prezi #jacksonbigdata

[Libby Lange @LangeLibby]

@claudiameng_ai true! although maybe stories also make disinfo more shareable: you only see each for a few seconds so not a lot of critical thinking goes into it, and then it disappears forever - difficult to track #jacksonbigdata

[Libby Lange @LangeLibby]

It seems like Instagram isn't discussed much when we talk about disinformation. I wonder why that is? #jacksonbigdata

[Libby Lange @LangeLibby]

I'm curious what portion of the population has the technical understanding to safely access the dark web - I certainly don't #jacksonbigdata

[Libby Lange @LangeLibby]

This talk really broadened the horizon of what feels possible with big data #jacksonbigdata

[Libby Lange @LangeLibby]

@sartorialdata wait what? please say more👀

[Libby Lange @LangeLibby]

The NSCAI Report recommended pretty much exactly this! A university geared towards AI research/tech that acts as a pipeline to public service #jacksonbigdata

[Libby Lange @LangeLibby]

I've never thought about the "energy efficiency" of the human brain! #jacksonbigdata

[Libby Lange @LangeLibby]

A good backgrounder on the group #jacksonbigdata:
https://t.co/lEHdCgYrUQ

[Libby Lange @LangeLibby]

This reminds me of the Jasons. Basically a group of scientists whose job was to advise the government on scientific possibilities, even if they were longshot projects. #jacksonbigdata

[Libby Lange @LangeLibby]

.@sartorialdata great point! The barrier to entry to being able to identify algorithmic bias is so high and the demographic with those skills is still too homogeneous #jacksonbigdata

[Libby Lange @LangeLibby]

@clairemcalkins I wonder if it's the sense of ubiquity. Whether or not this is accurate, the system in China has been painted as being based on everyday actions (did you cut in line at the coffee shop, what did you say about the Party in passing to someone else, etc.) #jacksonbigdata

[Libby Lange @LangeLibby]

Highly recommend the book Invisible Women! All about the data bias created when base assumptions were made with men in mind. #jacksonbigdata

[Libby Lange @LangeLibby]

I wonder what the market for personal datasets looks like (outside of social media companies). Are there companies that exist just to collect and sell these? #jacksonbigdata

[Libby Lange @LangeLibby]

I've had the same problem on my current project working with data @sartorialdata. You quickly realize you have to make a judgement about how comfortable you are with bridging the gap to your own conclusions. #jacksonbigdata

[Libby Lange @LangeLibby]

Correlation isn't enough when the data is still biased #jacksonbigdata

[Libby Lange @LangeLibby]

Dragonspeak is a great example of a good idea for big data's use - speech recognition - that falls short due to biases in its training data. Women's voices are misheard significantly more often. #jacksonbigdata