Deep Feature Space: A Geometrical Perspective

TL;DR

This paper introduces a novel visualization method for understanding the geometry of deep feature spaces in neural networks, focusing on overfitting and its impact on classification performance.

Contribution

It provides a new approach to visualize and analyze the intra-layer feature space, revealing insights into overfitting and improving recognition results.

Findings

Overfitting in feature space adversely affects classification.

The visualization method reveals intra-layer feature properties.

Proposed approach improves recognition accuracy.

Abstract

One of the most prominent attributes of Neural Networks (NNs) constitutes their capability of learning to extract robust and descriptive features from high dimensional data, like images. Hence, such an ability renders their exploitation as feature extractors particularly frequent in an abundant of modern reasoning systems. Their application scope mainly includes complex cascade tasks, like multi-modal recognition and deep Reinforcement Learning (RL). However, NNs induce implicit biases that are difficult to avoid or to deal with and are not met in traditional image descriptors. Moreover, the lack of knowledge for describing the intra-layer properties -- and thus their general behavior -- restricts the further applicability of the extracted features. With the paper at hand, a novel way of visualizing and understanding the vector space before the NNs' output layer is presented, aiming to enlighten the deep feature vectors' properties under classification tasks. Main attention is paid to the nature of overfitting in the feature space and its adverse effect on further exploitation. We present the findings that can be derived from our model's formulation, and we evaluate them on realistic recognition scenarios, proving its prominence by improving the obtained results.