HackerTube 中文?

USENIX Security '18-Q: Why Do Keynote Speakers Keep Suggesting That Improving Security Is Possible?

USENIXYouTube ↗

Further reading 1

How much does a successful demonstration establish? →A program gives the expected output. Between “it did” and “it will” lie the test distribution, formal assumptions and deployment context. Those boundaries are easy to miss under pressure to ship.This video helps answer · Does running a model explain its decision?

Caption notes

Mickens questions deploying opaque, hard-to-test models in high-stakes settings and argues that a security analysis must include social consequences.

Original passages · 6
you connect that inscrutable thing to important real life things like financial markets
USENIX Security '18-Q: Why Do Keynote Speakers Keep Suggesting That Improving Security Is Possible? · ▶ 16:27–16:37
how should that action be implemented to achieve societally beneficial outcomes
USENIX Security '18-Q: Why Do Keynote Speakers Keep Suggesting That Improving Security Is Possible? · ▶ 30:17–30:27
check your algorithms for biased input data to look for skews
USENIX Security '18-Q: Why Do Keynote Speakers Keep Suggesting That Improving Security Is Possible? · ▶ 50:06–50:13
deny by default firewall to prevent outsiders from probing those sensitive devices
USENIX Security '18-Q: Why Do Keynote Speakers Keep Suggesting That Improving Security Is Possible? · ▶ 45:24–45:33
systems do the right thing even in the presence of malicious inputs
USENIX Security '18-Q: Why Do Keynote Speakers Keep Suggesting That Improving Security Is Possible? · ▶ 5:27–5:34
technology is value neutral and will therefore automatically lead to good outcomes
USENIX Security '18-Q: Why Do Keynote Speakers Keep Suggesting That Improving Security Is Possible? · ▶ 37:02–37:10

Hacker News comments

View original on HN ↗