# Differentiable Algorithm Networks for Composable Robot Learning

## TL;DR

This paper presents Differentiable Algorithm Networks (DAN), a modular, end-to-end trainable architecture that combines model-driven and data-driven approaches for robot learning, demonstrated on a navigation task in complex environments.

## Contribution

The paper introduces DAN, a novel architecture that integrates differentiable robot algorithms with neural modules, enabling efficient learning and adaptation in robotic systems.

## Key findings

- DAN effectively learns navigation in complex 3D environments.
- Modules can adapt to imperfect models and algorithms.
- End-to-end training improves overall system performance.

## Abstract

This paper introduces the Differentiable Algorithm Network (DAN), a composable architecture for robot learning systems. A DAN is composed of neural network modules, each encoding a differentiable robot algorithm and an associated model; and it is trained end-to-end from data. DAN combines the strengths of model-driven modular system design and data-driven end-to-end learning. The algorithms and models act as structural assumptions to reduce the data requirements for learning; end-to-end learning allows the modules to adapt to one another and compensate for imperfect models and algorithms, in order to achieve the best overall system performance. We illustrate the DAN methodology through a case study on a simulated robot system, which learns to navigate in complex 3-D environments with only local visual observations and an image of a partially correct 2-D floor map.