Learning the Space of Deep Models
- URL: http://arxiv.org/abs/2206.05194v1
- Date: Fri, 10 Jun 2022 15:53:35 GMT
- Title: Learning the Space of Deep Models
- Authors: Gianluca Berardi, Luca De Luigi, Samuele Salti, Luigi Di Stefano
- Abstract summary: We show how it is possible to use representation learning to learn a fixed-size, low-dimensional embedding space of trained deep models.
We address image classification and neural representation of signals, showing how our embedding space can be learnt so as to capture the notions of performance and 3D shape.
- Score: 12.785442275745973
- License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
- Abstract: Embedding of large but redundant data, such as images or text, in a hierarchy
of lower-dimensional spaces is one of the key features of representation
learning approaches, which nowadays provide state-of-the-art solutions to
problems once believed hard or impossible to solve. In this work, in a plot
twist with a strong meta aftertaste, we show how trained deep models are as
redundant as the data they are optimized to process, and how it is therefore
possible to use deep learning models to embed deep learning models. In
particular, we show that it is possible to use representation learning to learn
a fixed-size, low-dimensional embedding space of trained deep models and that
such space can be explored by interpolation or optimization to attain
ready-to-use models. We find that it is possible to learn an embedding space of
multiple instances of the same architecture and of multiple architectures. We
address image classification and neural representation of signals, showing how
our embedding space can be learnt so as to capture the notions of performance
and 3D shape, respectively. In the Multi-Architecture setting we also show how
an embedding trained only on a subset of architectures can learn to generate
already-trained instances of architectures it never sees instantiated at
training time.
Related papers
- Training Neural Networks with Internal State, Unconstrained
Connectivity, and Discrete Activations [66.53734987585244]
True intelligence may require the ability of a machine learning model to manage internal state.
We show that we have not yet discovered the most effective algorithms for training such models.
We present one attempt to design such a training algorithm, applied to an architecture with binary activations and only a single matrix of weights.
arXiv Detail & Related papers (2023-12-22T01:19:08Z) - Breaking the Curse of Dimensionality in Deep Neural Networks by Learning
Invariant Representations [1.9580473532948401]
This thesis explores the theoretical foundations of deep learning by studying the relationship between the architecture of these models and the inherent structures found within the data they process.
We ask What drives the efficacy of deep learning algorithms and allows them to beat the so-called curse of dimensionality.
Our methodology takes an empirical approach to deep learning, combining experimental studies with physics-inspired toy models.
arXiv Detail & Related papers (2023-10-24T19:50:41Z) - Multi-domain learning CNN model for microscopy image classification [3.2835754110596236]
We present a multi-domain learning architecture for the classification of microscopy images.
Unlike previous methods that are computationally intensive, we have developed a compact model, called Mobincep.
It surpasses state-of-the-art results and is robust for limited labeled data.
arXiv Detail & Related papers (2023-04-20T19:32:23Z) - ROAD: Learning an Implicit Recursive Octree Auto-Decoder to Efficiently
Encode 3D Shapes [32.267066838654834]
We present a novel implicit representation to efficiently and accurately encode large datasets of complex 3D shapes.
Our implicit Recursive Octree Auto-Decoder (ROAD) learns a hierarchically structured latent space enabling state-of-the-art reconstruction results at a compression ratio above 99%.
arXiv Detail & Related papers (2022-12-12T19:09:47Z) - Contrastive Neighborhood Alignment [81.65103777329874]
We present Contrastive Neighborhood Alignment (CNA), a manifold learning approach to maintain the topology of learned features.
The target model aims to mimic the local structure of the source representation space using a contrastive loss.
CNA is illustrated in three scenarios: manifold learning, where the model maintains the local topology of the original data in a dimension-reduced space; model distillation, where a small student model is trained to mimic a larger teacher; and legacy model update, where an older model is replaced by a more powerful one.
arXiv Detail & Related papers (2022-01-06T04:58:31Z) - LOLNeRF: Learn from One Look [22.771493686755544]
We present a method for learning a generative 3D model based on neural radiance fields.
We show that, unlike existing methods, one does not need multi-view data to achieve this goal.
arXiv Detail & Related papers (2021-11-19T01:20:01Z) - A Design Space Study for LISTA and Beyond [79.76740811464597]
In recent years, great success has been witnessed in building problem-specific deep networks from unrolling iterative algorithms.
This paper revisits the role of unrolling as a design approach for deep networks, to what extent its resulting special architecture is superior, and can we find better?
Using LISTA for sparse recovery as a representative example, we conduct the first thorough design space study for the unrolled models.
arXiv Detail & Related papers (2021-04-08T23:01:52Z) - S2RMs: Spatially Structured Recurrent Modules [105.0377129434636]
We take a step towards exploiting dynamic structure that are capable of simultaneously exploiting both modular andtemporal structures.
We find our models to be robust to the number of available views and better capable of generalization to novel tasks without additional training.
arXiv Detail & Related papers (2020-07-13T17:44:30Z) - Mutual Information Maximization for Robust Plannable Representations [82.83676853746742]
We present MIRO, an information theoretic representational learning algorithm for model-based reinforcement learning.
We show that our approach is more robust than reconstruction objectives in the presence of distractors and cluttered scenes.
arXiv Detail & Related papers (2020-05-16T21:58:47Z) - Learning Deformable Image Registration from Optimization: Perspective,
Modules, Bilevel Training and Beyond [62.730497582218284]
We develop a new deep learning based framework to optimize a diffeomorphic model via multi-scale propagation.
We conduct two groups of image registration experiments on 3D volume datasets including image-to-atlas registration on brain MRI data and image-to-image registration on liver CT data.
arXiv Detail & Related papers (2020-04-30T03:23:45Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.