IMO, most AI based malware detection companies stand on shoulder of ONE giant, which is @virustotal.
This reminds me, when few years ago, VT changed its contribution policy under pressure of legacy AV vendors whose detections being used by new AI vendors to train their models.
VirusTotal is arguably the ImageNet of cybersecurity AI, having provided large-scale labeled data and a rough means by which to compare approaches. This was a happy accident though, and is not VT's purpose -- which is why efforts like @mrphilroth et al.'s EMBER are so important.https://twitter.com/ChristophSchll3/status/1219360589920841740 …
-
-
-
+1, agree about many companies' reliance on VirusTotal, particularly startups. Feels complicated though... some of those startups made good, open contributions in the AI/security space thanks to VT, which legacy vendors then adopted. And some were probably less than ethical :/
Kraj razgovora
Novi razgovor -
-
-
Nice. I gave enough talks about EMBER to rise out of "et al." status and push you back down into it. Take that
@drhyrum! -
Oh man, I didn't know what the decorum was, and I ran out of characters in my tweet, hence et al :/
- Još 1 odgovor
Novi razgovor -
-
-
I hope it morphs into more then a labeling engine - benchmarking Security specific ML models aka products
-
One can hope... the issue is vendors would pull out of the platform if they provided per-vendor accuracy measures. Many of us in the MLsec vendor community do our own assessments of our engines' accuracy versus our colleagues', but we can't share these externally.
- Još 1 odgovor
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.