LK-99 Updates | LLM Editing | AI Radiology Study

08/02/23 • 15 min

In this episode of AI Daily, hosts Farb, Ethan, and Conner delve into three big stories in the world of AI. First, discover the ripple effects of knowledge editing in language models, a benchmark of 5,000 facts highlighting challenges in current LLM editing, and an innovative in-context editing method. Next, we bring you updates on LK-99, a room temperature superconductor that may revolutionize the field. Learn about simulation findings and the potential end of Wakanda's unobtainium monopoly. Lastly, we explore how AI is impacting the field of Radiology. Uncover whether AI copilots or working independently is more effective for radiologists and the role of UX in AI adoption.

Quick Points

1️⃣ LLM Editing

Adding or changing a single fact can cause a cascade of changes in an LLM's understanding

Benchmark of 5,000 facts reveals current LLM editing methods struggle with ripple effects.

Innovative in-context editing method shows promising results.

2️⃣ LK-99 Updates

LK-99 superconductor shows potential with simulated copper bands for energy transfer.

Exciting news shifts markets as room temperature superconductivity gains traction.

Future engineering may lead to increased bands for practical superconducting applications.

3️⃣ AI Radiology Study

Combining AI and human expertise in radiology yields suboptimal results.

UX plays a vital role in AI adoption for medical applications.

Future implications suggest AI or human-only approaches may be more effective.

🔗 Episode Links

LLM Editing Paper

LK-99 Updates Tweet #1

LK-99 Updates Tweet #2

AI Radiology Study

Neon Series B

GPU Supply & Demand

Connect With Us:

Subscribe to our Substack

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.aidailypod.com

Quick Points

1️⃣ LLM Editing

Adding or changing a single fact can cause a cascade of changes in an LLM's understanding

Benchmark of 5,000 facts reveals current LLM editing methods struggle with ripple effects.

Innovative in-context editing method shows promising results.

2️⃣ LK-99 Updates

LK-99 superconductor shows potential with simulated copper bands for energy transfer.

Exciting news shifts markets as room temperature superconductivity gains traction.

Future engineering may lead to increased bands for practical superconducting applications.

3️⃣ AI Radiology Study

Combining AI and human expertise in radiology yields suboptimal results.

UX plays a vital role in AI adoption for medical applications.

Future implications suggest AI or human-only approaches may be more effective.

🔗 Episode Links

LLM Editing Paper

LK-99 Updates Tweet #1

LK-99 Updates Tweet #2

AI Radiology Study

Neon Series B

GPU Supply & Demand

Connect With Us:

Subscribe to our Substack

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.aidailypod.com

Previous Episode

Meta's OpenCatalyst | RT-2 Speaking Robot | Adversarial Prompts

In this episode of AI Daily with your hosts Conner, Ethan, and Farb. They kick off the episode discussing Meta's OpenCatalyst, a groundbreaking model developed with Carnegie Mellon University that simulates over a hundred million catalyst combinations, accelerating advancements in material science and renewable energy. They then move to explore Google DeepMind's RT-2 Speaking Robot, a unique vision, language, and action model that learns from web images and texts to perform real-world actions, promising a new era of autonomous robotics. Finally, they delve into the intriguing concept of Adversarial Prompts, discussing a recent study by a team at Carnegie Mellon that used LLaMA to generate prompts adversarial to popular models like GPT-4, raising important questions about the robustness and safety of these models.

Quick Points:

1️⃣ Meta’s OpenCatalyst

Meta and Carnegie Mellon University develop OpenCatalyst, simulating 100+ million catalyst combinations.

This tool enables rapid simulations, enhancing chemical process research.

It is highly applicable to renewable energy and material sciences.

2️⃣ RT-2 Speaking Robot

Google DeepMind unveils the RT-2 Speaking Robot, a vision-language-action model.

Trained on web images and texts, it can perform untrained real-world actions.

This model represents a significant leap in the realm of autonomous robotics.

3️⃣ Adversarial Prompts

A Carnegie Mellon team uses LLaMA to generate adversarial prompts against leading models.

This discovery exposes potential weaknesses in popular AI models like GPT-4.

Raises important questions about AI model robustness and safety.

🔗 Episode Links

Connect With Us:

Subscribe to our Substack

This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit www.aidailypod.com

Next Episode

HierVST Voice Cloning | NVIDIA Perfusion | Meta's AudioCraft

Welcome to AI Daily! Join hosts Farb, Ethan, and Conner as they explore three groundbreaking AI stories First up, HierVST Voice Cloning - Experience zero-shot voice cloning with impressive accuracy using just one audio clip. Next, NVIDIA Perfusion - a small, powerful personalization model for text images, using key locking to maintain consistency. Lastly, Meta's AudioCraft - the fusion of music generation, audio generation, and codecs into one open-source code base, creating high-fidelity outputs.

Quick Points

1️⃣ HierVST Voice Cloning

Zero-shot voice cloning system achieves accurate outputs with just one audio clip.

Uses hierarchical models for long and short-term generation understanding.

Potential challenges in handling longer clips and need for further fine-tuning.

2️⃣ NVIDIA Perfusion

Personalization model for text images with key locking for subject consistency.

Only 100 kilobytes, trains in four minutes, and outperforms other models.

Open-source codebase, but may need improvements for human subjects.

3️⃣ Meta’s AudioCraft

Audio generation, music gen, and codecs combined into an open-source codebase.

High-fidelity outputs, 30 seconds of sounds, compressing audio files efficiently.