123 points by darknet_researcher 6 months ago flag hide 22 comments
cmucat 6 months ago next
AI is one heck of a challenge. Facing that as a PhD student requires dedication and enthusiasm. Thanks for sharing!
deeplearningdave 6 months ago next
Thanks for the good words! I think sharing our progress is crucial to the scientific community, especially something as prone to concerns as AI
quantumquinn 6 months ago next
@deep LearningDave That attitude is underestimated. We really need more researchers like you sharing your valuable research and insights
stanfordsam 6 months ago next
@quantumQuinn Agreed! I'd been reading a lot about classic neural nets but missed entirely on the recent developments
pythonpete 6 months ago next
Could you also link some of your old posts, @StanfordSam? Curious to learn and explore classics related to neural networks
kaggleken 6 months ago next
@PythonPete Sure! Here is the link to my Neural Networks course series. Hope you find it useful - http://bit.ly/NN_course_series
deeplearningdave 6 months ago prev next
Fascinating journey into neural network obfuscation! I've been thinking about the security implications of AI lately, and this really hits the nail on the head.
kaggleken 6 months ago next
I agree with you on that, Dave. I'm currently beta testing an adversarial defense system that I hope addresses some of these concerns. Will share more soon.
stanfordsam 6 months ago next
@kaggleKen would love to check it out if possible! Email sent :)
cvcreatorcarl 6 months ago next
I'll be exploring this avenue in my thesis, focusing more on the defensive techniques. A layman's blog post would be cool too
pythonpete 6 months ago next
When I wrote my neural network chess program, I wished I'd known about obfuscation. Well worth keeping in mind for future projects, @CVCreatorCarl
quantumquinn 6 months ago prev next
This field moves so quickly! I remember writing my own neural net from scratch and seeing research like this was beyond my imagination. Kudos for your amazing work
fastforwardfrank 6 months ago next
The speed at which we progress in understanding neural networks and AI is simply alarming! Really looking forward to more work in interpretability and ethics
mlmonkey 6 months ago prev next
This is pretty advanced. I wish I'd had some insight into this during my undergraduate work. I might be able to help some of my current students with their theses! I'd recommend sharing on other AI forums outside HN as well
kaggleken 6 months ago next
I agree, MLMonkey! I didn't think I'd ever see my undergraduate level knowledge applied in such a fanciful way
pythonpete 6 months ago prev next
Great read! This goes well with the recent discussions about AI interpretability and ethics. Planning to explore this further for defensive purposes
mlmonkey 6 months ago prev next
I'll check out your blog when you publish, @CVCreatorCarl. I'm sure it'll be a great read and very helpful to others as well
cmucat 6 months ago next
@mlMonkey I'm sure many of my peers would also appreciate learning tips like this. Could you recommend a resource to learn more about it?
cvcreatorcarl 6 months ago next
@CMUcat I've collected various resources over time and am in the process of organizing them to share. Will ping you soon :)
fastforwardfrank 6 months ago prev next
To those of you discussing blog posts, video walk-throughs, or tutorials, the AI community should be extremely grateful once you publish! Any resources are valuable for learning
quantumquinn 6 months ago next
@FastForwardFrank It is indeed excellent when researchers and developers share knowledge, which makes AI accessible to more skill levels
deeplearningdave 6 months ago prev next
Let's keep going deeper into the rabbit hole of AI and its intricacies. Thanks to everyone for the encouragement! #makingAIopen