S2Lab

Official Twitter account of the Systems Security Research Lab, in the Information Security Research Group at UCL Computer Science

[Konrad Rieck 🌈 @mlsec]

Did security folks not value your ML insights? Were ML reviewers unimpressed by your security analysis? Maybe you’re submitting to the wrong conference.

@satml_conf has you covered! Our Call for Papers is open now!

👉 https://t.co/3zPzuWI3Ax
⏰ Deadline: Sep 18

[S2Lab @s2lab]

@fbpierazzi Congratulations Fabio! :-)

[Lorenzo Cavallaro @lcavallaro]

Good news, everyone! (cit.) 2x brand new @uclisec faculty positions @uclcs @ucl - application deadline in November.

Join us to make that little dents in the knowledge we're all aiming for, embracing diversity and inclusion at its core, for a better society.

w00t w00t!

[Ambra @ambrademontis]

Considering the current uncertainty and limitations, we have rescheduled AISec https://t.co/NbZMhPg3YV as a virtual event and postponed the paper submission deadline to July 26.
@acm_ccs #MLsec #MachineLearning #AI #adversarial #malware @adversarial_ML @trustworthy_ml

[Lorenzo Cavallaro @lcavallaro]

Excited to share recent advances or thought-provoking research in #AISec? Look no further and submit your work to https://t.co/3rJXUVtW3V co-located with @acm_ccs #CCS21 - deadline 6/29/2021 approaching 😱😎

[Lorenzo Cavallaro @lcavallaro]

Did you miss Deep Learning and Security @IEEESSP 2021? Longing for reliving it? Worry no more!

We. Has. Recordings:

DLS2021 Pt1/2: https://t.co/GBrLzb0nd6
DLS2021 Pt2/2: https://t.co/AAJ3xQYx63

Keynotes, papers, and an exciting and thought-provoking panel. The full package.

[Lorenzo Cavallaro @lcavallaro]

Truly honored to talk about #AdversarialML in the context of realizable attacks and the interesting role program analysis and machine learning play. Thanks @CyberSecAI also for the opportunity to share additional thoughts on the topic! @kclinformatics @KingsNMS @KingsCollegeLon

[@[email protected] @barsteward]

@BSidesLondon @SecurityBSides @mikko’s 2018 Keynote https://t.co/DsPpCECash

@ghostie_ on physical pentests
https://t.co/58TmoAO8fA

@lcavallaro on flaws in ML https://t.co/PBHfS0YWvD

@jessicaamery_ on stalkerware https://t.co/alQJYqtoKj

#BSidesLDN2020 #PandemicPostponed #SeeYouInOctober

[Lorenzo Cavallaro @lcavallaro]

It’s been an honor to serve as a Program Co-Chair with @lanziandrea to @EuroSecWorkshop. Despite the challenging time, we are “So close, no matter how far; Couldn't be much more from the heart; Forever trusting who we are; And nothing else matters” (cit.) Great keynote & talks!

[Gergely Neu @neu_rips]

the picture i use every time i'm teaching dynamic programming
(source unknown)

[S2Lab @s2lab]

The trailer of “Intriguing Properties of Adversarial ML Attacks in the Problem Space”, our @IEEESSP #SP20 work, is live! https://t.co/KGL1jlk464

Our reformulation of #adversarial #ML in the problem space shows evading detection is a realistic threat even in constrained domains

[DIMVA Conference @DIMVAConf]

Considering the COVID-19 situation, a decision has been made to make DIMVA2020 a video-based digital conference. The dates will be the same and proceedings will be published as planned. Registration will not be required to attend the conference. More details will be given soon.

[Megan McCurry @microbe_megan]

It’s always Reviewer 3 🤦🏽‍♀️ This is for everyone working on manuscripts during the #COVID19 lab shutdown 🙏🏼 #phdchat #AcademicChatter

[Lorenzo Cavallaro @lcavallaro]

The list of accepted papers to EuroSec 2020 is out! Interesting research that will spark lots of discussions - join us online on Apr 27! Stay tuned for further details :-)

[Prof. Akiko Iwasaki @VirusesImmunity]

In this new piece, I present "Antidote to toxic principal investigators". We must protect our next generation scientists from the principal investigators who abuse their powers and mistreat their trainees. (This piece focuses on graduate students)

https://t.co/6IMilcfCwN

[Brendan Dolan-Gavitt @moyix]

DIMVA is one of my favorites; it punches above its weight in terms of paper quality. Definitely jump on the opportunity to virtually attend!

[S2Lab @s2lab]

We have to limit our physical interactions, but we can enhance our virtual ones: here's the new slick Systems Security Research Lab website to keep us closer (kudos to @Entrophy6 and @0xtrustypatches for bringing this to life): https://t.co/3BilBP58Gf

[Brendan Dolan-Gavitt @moyix]

Temporal biases in malware ML evaluations (e.g., training on data that appeared later than the test data) are endemic in the published literature! Interesting work from @fbpierazzi

[Lorenzo Cavallaro @lcavallaro]

@fbpierazzi at #usesec19 tells us about our work on time-space experimental bias and the effect of concept drift in malware classifiers: TESSERACT - towards sound and realistic evaluations & comparisons of ML for security tasks: https://t.co/GYrQQyYqsO

[Lorenzo Cavallaro @lcavallaro]

Evasive attacks on ML classifiers are scary but we know little about adversarial ML in the problem space: we may need to rethink the whole formalization, incl. optimization and constraints for a principled approach - come and talk to us @USENIXSecurity poster session #usesec19