-
Variational Inference
Mathematics of Evidence Lower Bound (ELBO) objective formulation is inevitably encountered when dealing with latent dynamics models and VAEs. In this blog, we break it down along with necessary concepts like...
-
Algorithmic Thinking
Nature is often super secretive in her approaches. Yet sometimes she unveils her way of making things work cleverly. She works around complicated problems with as minimal computing power as possible still ma...
-
World Models
Advantage of learning a world model is that we can make agents learn in the simulated environment generated by the world model itself and develop a policy which can later be used in the actual environment...
-
Temporal Difference Contrastive GCRL
To capture long-term temporal dependency: larger quantities of data is needed. In this blog, we will talk about ways to make it more sample efficient using temporal difference version of ranking-based contrast...
-
Latent Dynamics
Latent dynamics describe how an agent models the evolution of its environment in a latent space—a compact, abstract representation of high-dimensional observations. Unlike direct interaction with raw data, ...
-
Exploring Wormhole With Quantum Computer
Just like how from a 2D pattern, we can project out a 3D hologram, it says from n-1 dimensional quantum system, emergence of n dimensional space-time can happen. Since the 1930s, physicists have tried to ...
-
Quantum Computing
How tough do you think is it to simulate a single molecule of caffeine? Around 10⁴⁸ bits are required to represent it accounting for protons, neutrons, electrons, energy configurations, and all of its bonds...
-
Delusion
It was dark, pitch dark as if the light was forbidden in this reign of the void. “My body?” — a sense of simultaneous squeezing and stretching was there just a few nanoseconds before. After the spaghettification, it was just consciousness that remained. “Was it….DEATH?”