Категории

Nightshade: a poison pill that subtly distorts image composition, causing model confusion & collapse

Опубликовано: 08 Октябрь 2024
на канале: The TWIML AI Podcast with Sam Charrington
743
19

Nightshade: Data Poisoning to Fight Generative AI with Ben Zhao - Full Interview: https://twimlai.com/go/668

🔔 Subscribe to our channel for more great ML content just like this: https://youtube.com/twimlai?sub_confi...

🗣️ CHAT WITH US!
Subscribe to the TWIML AI Podcast: https://twimlai.com/podcast/twimlai/
Join our Slack Community: https://twimlai.com/community/
Subscribe to our newsletter: https://twimlai.com/newsletter/
Want to get in touch? Send us a message: https://twimlai.com/contact/

#ai #machinelearning #generativeai

play_arrow
625,844
6.3 тыс

Do you have Nutella and puff pastry at home? You will NEVER buy croissants again!

Do you have Nutella and puff pastry at home? You will NEVER buy croissants again!
play_arrow
87,006
2.5 тыс

SUNDAY NIGHT INSPECTION ON OME TV

SUNDAY NIGHT INSPECTION ON OME TV
play_arrow
112
4

Современный медицинский центр АРЕНСИЯ

Современный медицинский центр АРЕНСИЯ
play_arrow
2,327
59

Lemme lock in 😤😤😤

Lemme lock in 😤😤😤

play_arrow
568
4

Post and Page Builder - Working with Images

Post and Page Builder - Working with Images
play_arrow
122,375
13 тыс

FUI PEDIDA EM NAMORO!!!

FUI PEDIDA EM NAMORO!!!
play_arrow
311,861
3.3 тыс

CNL - Mulan Jameela - Makhluk Tuhan Paling Sexy

CNL - Mulan Jameela - Makhluk Tuhan Paling Sexy
play_arrow
379
11

Dirty Job, Mike Campbell & the Dirty Knobs, The Guild Theater, Menlo Park, CA 5-19-2023

Dirty Job, Mike Campbell & the Dirty Knobs, The Guild Theater, Menlo Park, CA 5-19-2023
Похожие видео
play_arrow
Fatih shares video language model providing real-time feedback and coaching on activities #CVPR2024

Fatih shares video language model providing real-time feedback and coaching on activities #CVPR2024
play_arrow
Do LLMs and RL offer a path to AGI? #machinelearning #ai #podcast

Do LLMs and RL offer a path to AGI? #machinelearning #ai #podcast
play_arrow
How risky are open-source LLMs really? #machinelearning #ai #podcast

How risky are open-source LLMs really? #machinelearning #ai #podcast
play_arrow
Video generators are world models #machinelearning #ai #podcast

Video generators are world models #machinelearning #ai #podcast
play_arrow
Are GPUs enough to run AGI? #ai #podcast #machinelearning #reinforcementlearning

Are GPUs enough to run AGI? #ai #podcast #machinelearning #reinforcementlearning
play_arrow
Nightshade: a poison pill that subtly distorts image composition, causing model confusion & collapse

Nightshade: a poison pill that subtly distorts image composition, causing model confusion & collapse
play_arrow
The main challenge comes from difficulties in discrete optimization. Scaling needs improved strategy

The main challenge comes from difficulties in discrete optimization. Scaling needs improved strategy
play_arrow
Prediction of shift in LLMs: Embedding desiderata in training to avoid post hoc interventions.

Prediction of shift in LLMs: Embedding desiderata in training to avoid post hoc interventions.
play_arrow
Success lies in listening, engaging, adapting to evolving needs and building sticky, useful products

Success lies in listening, engaging, adapting to evolving needs and building sticky, useful products
play_arrow
Retrieval patterns allows the model to act as a ‘problem-solving brain’, yielding smarter solutions.

Retrieval patterns allows the model to act as a ‘problem-solving brain’, yielding smarter solutions.
play_arrow
Turn problems to checklists and restructure from low to high-level judgments to avoid hallucinations

Turn problems to checklists and restructure from low to high-level judgments to avoid hallucinations
play_arrow
Evaluating AI's general performance & finding models for unrestricted prompting remains a challenge

Evaluating AI's general performance & finding models for unrestricted prompting remains a challenge
play_arrow
AI shifts from engineering to complex system science, struggles with understanding emergent behavior

AI shifts from engineering to complex system science, struggles with understanding emergent behavior
play_arrow
Altering the weights is vital in adding concepts. Hypernetwork personalize weights for faster 🖼️ gen

Altering the weights is vital in adding concepts. Hypernetwork personalize weights for faster 🖼️ gen
play_arrow
The ML metrics are behind where the tech is. LLMs score lower on metrics but generate better results

The ML metrics are behind where the tech is. LLMs score lower on metrics but generate better results
play_arrow
"LLMs are narcissistic in that they find their own work better than a human-generated text” - Shreya

"LLMs are narcissistic in that they find their own work better than a human-generated text” - Shreya
play_arrow
Is RAG an essential component of the LLM stack? #llm #aisafety #artificialintelligence

Is RAG an essential component of the LLM stack? #llm #aisafety #artificialintelligence
play_arrow
Language 🗣️ is the key ingredient to achieving human-like AI 🧠

Language 🗣️ is the key ingredient to achieving human-like AI 🧠
play_arrow
Consistency & robustness in #generativeai models are crucial for software stability and performance

Consistency & robustness in #generativeai models are crucial for software stability and performance
play_arrow
Just how different were the results between GPT-3.5 and GPT-4?

Just how different were the results between GPT-3.5 and GPT-4?
play_arrow
Efficiency: A fundamental principle in neuroscience

Efficiency: A fundamental principle in neuroscience
play_arrow
Teaching Large Language Models to Stick to the Script

Teaching Large Language Models to Stick to the Script
play_arrow
AI Ethics; A practioners perspective

AI Ethics; A practioners perspective
play_arrow
Causal Reasoning in Machine Learning

Causal Reasoning in Machine Learning