Sign in Subscribe

aimodels-fyi

Synthesized hand wave

They found a new NeRF technique to turn videos into controllable 3D models

Creating realistic, animated 3D models from video footage has been a longstanding challenge in the field of computer graphics due to the complexity of human movement and the subtleties of appearance under varying conditions. Traditionally, this process has relied on costly and labor-intensive techniques such as multi-camera setups and detailed

PPLs deal with probability

Diffusion might be a better way to model probability in PPLs

DMVI uses diffusion models to approximate the probability distributions for faster, more accurate automated inference.

How FACTOR works

FACTOR: Detecting Deepfakes without Seeing Any Examples

As deepfake technology advances rapidly, we need better ways to detect them. Most current methods train on available deepfakes, but fail to generalize to new, unseen ones.

Example results

Anomaly Detection in Multivariate Time Series with... Diffusion Models?

Diffusion processes are great at smoothing out normal patterns while amplifying anomalies - perfect for AD.

Emotional Prompting improves AI preformance

Telling GPT-4 you're scared or under pressure improves performance

Researchers show LLMs respond with improved performance when prompted with emotional context

RT-2-X image from the paper

DeepMind's X-Embodiment and RT-X: Training Robots to Learn from Each Other

Just as people acquire new skills by observing others, robots also benefit from pooling knowledge

Example implementation of PubDef and transfer attack

PubDef: Defending Against Transfer Attacks Using Public Models

Adversarial attacks pose a serious threat to ML models. But most proposed defenses hurt performance on clean data too much to be practical.