Structured Memory based Deep Model to Detect as well as Characterize Novel Inputs

TL;DR

This paper introduces a novel deep learning architecture with structured memory that can detect and characterize new, unseen inputs by relating them to memorized past representations, enhancing reasoning capabilities.

Contribution

It proposes a spatially forked deep model with a structured memory bank and a comparative learner for improved detection and characterization of novel inputs.

Findings

Successfully characterized unseen object classes in synthetic datasets

Demonstrated ability to relate new inputs to memorized representations

Enhanced reasoning about input nature beyond training data

Abstract

While deep learning has pushed the boundaries in various machine learning tasks, the current models are still far away from replicating many functions that a normal human brain can do. Explicit memorization based deep architecture have been recently proposed with the objective to understand and predict better. In this work, we design a system that involves a primary learner and an adjacent representational memory bank which is organized using a comparative learner. This spatially forked deep architecture with a structured memory can simultaneously predict and reason about the nature of an input, which may even belong to a category never seen in the training data, by relating it with the memorized past representations at the higher layers. Characterizing images of unseen object classes in both synthetic and real world datasets is used as an example to showcase the operational success of the proposed framework.