English
Related papers

Related papers: Learning Features with Parameter-Free Layers

200 papers

In the pursuit of reducing the number of trainable parameters in deep transformer networks, we employ Reinforcement Learning to dynamically select layers during training and tie them together. Every few iterations, the RL agent is asked…

Machine Learning · Computer Science 2024-01-24 Tamir David Hay , Lior Wolf

In response to the development of recent efficient dense layers, this paper shows that something as simple as replacing linear components in pointwise convolutions with structured linear decompositions also produces substantial gains in the…

Machine Learning · Statistics 2019-06-04 Gavin Gray , Elliot J. Crowley , Amos Storkey

Modern deep neural networks are typically highly overparameterized. Pruning techniques are able to remove a significant fraction of network parameters with little loss in accuracy. Recently, techniques based on dynamic reallocation of…

Machine Learning · Computer Science 2019-05-14 Hesham Mostafa , Xin Wang

Averaging neural network parameters is an intuitive method for fusing the knowledge of two independent models. It is most prominently used in federated learning. If models are averaged at the end of training, this can only lead to a good…

Machine Learning · Computer Science 2024-03-20 Linara Adilova , Maksym Andriushchenko , Michael Kamp , Asja Fischer , Martin Jaggi

In this paper we derive an efficient algorithm to learn the parameters of structured predictors in general graphical models. This algorithm blends the learning and inference tasks, which results in a significant speedup over traditional…

Machine Learning · Computer Science 2013-09-02 Tamir Hazan , Alexander Schwing , David McAllester , Raquel Urtasun

Behavior of neural networks is irremediably determined by the specific loss and data used during training. However it is often desirable to tune the model at inference time based on external factors such as preferences of the user or…

Computer Vision and Pattern Recognition · Computer Science 2023-04-04 Matteo Maggioni , Thomas Tanay , Francesca Babiloni , Steven McDonagh , Aleš Leonardis

The empirical success of deep learning is often attributed to deep networks' ability to exploit hierarchical structure in data, constructing increasingly complex features across layers. Yet despite substantial progress in deep learning…

Machine Learning · Computer Science 2026-01-28 Yunwei Ren , Yatin Dandi , Florent Krzakala , Jason D. Lee

The convolution operation is a central building block of neural network architectures widely used in computer vision. The size of the convolution kernels determines both the expressiveness of convolutional neural networks (CNN), as well as…

Image and Video Processing · Electrical Eng. & Systems 2022-10-10 Tianyu Ma , Adrian V. Dalca , Mert R. Sabuncu

Filters in convolutional networks are typically parameterized in a pixel basis, that does not take prior knowledge about the visual world into account. We investigate the generalized notion of frames designed with image properties in mind,…

Computer Vision and Pattern Recognition · Computer Science 2017-07-20 Jörn-Henrik Jacobsen , Bert de Brabandere , Arnold W. M. Smeulders

We consider layerwise function-space learning rates, which measure the magnitude of the change in a neural network's output function in response to an update to a parameter tensor. This contrasts with traditional learning rates, which…

Machine Learning · Statistics 2025-05-23 Edward Milsom , Ben Anson , Laurence Aitchison

Convolutional Neural Networks (CNNs) have achieved state-of-the-art performance in many computer vision tasks over the years. However, this comes at the cost of heavy computation and memory intensive network designs, suggesting potential…

Computer Vision and Pattern Recognition · Computer Science 2020-08-11 Kumara Kahatapitiya , Ranga Rodrigo

State-of-the-art rehearsal-free continual learning methods exploit the peculiarities of Vision Transformers to learn task-specific prompts, drastically reducing catastrophic forgetting. However, there is a tradeoff between the number of…

Computer Vision and Pattern Recognition · Computer Science 2023-08-21 Thomas De Min , Massimiliano Mancini , Karteek Alahari , Xavier Alameda-Pineda , Elisa Ricci

Convolutional Neural Networks (CNNs) do not have a predictable recognition behavior with respect to the input resolution change. This prevents the feasibility of deployment on different input image resolutions for a specific model. To…

Computer Vision and Pattern Recognition · Computer Science 2020-07-14 Duo Li , Anbang Yao , Qifeng Chen

We develop an algorithm for systematic design of a large artificial neural network using a progression property. We find that some non-linear functions, such as the rectifier linear unit and its derivatives, hold the property. The…

Neural and Evolutionary Computing · Computer Science 2017-10-24 Saikat Chatterjee , Alireza M. Javid , Mostafa Sadeghi , Partha P. Mitra , Mikael Skoglund

Many different deep networks have been used to approximate, accelerate or improve traditional image operators. Among these traditional operators, many contain parameters which need to be tweaked to obtain the satisfactory results, which we…

Computer Vision and Pattern Recognition · Computer Science 2019-07-15 Qingnan Fan , Dongdong Chen , Lu Yuan , Gang Hua , Nenghai Yu , Baoquan Chen

The large-scale visual pretraining has significantly improve the performance of large vision models. However, we observe the \emph{low FLOPs pitfall} that the existing low-FLOPs models cannot benefit from large-scale pretraining. In this…

Computer Vision and Pattern Recognition · Computer Science 2024-01-17 Kai Han , Yunhe Wang , Jianyuan Guo , Enhua Wu

We present flattened convolutional neural networks that are designed for fast feedforward execution. The redundancy of the parameters, especially weights of the convolutional filters in convolutional neural networks has been extensively…

Neural and Evolutionary Computing · Computer Science 2015-11-23 Jonghoon Jin , Aysegul Dundar , Eugenio Culurciello

Memory layers use a trainable key-value lookup mechanism to add extra parameters to a model without increasing FLOPs. Conceptually, sparsely activated memory layers complement compute-heavy dense feed-forward layers, providing dedicated…

Computation and Language · Computer Science 2024-12-23 Vincent-Pierre Berges , Barlas Oğuz , Daniel Haziza , Wen-tau Yih , Luke Zettlemoyer , Gargi Ghosh

Convolutional layers are one of the basic building blocks of modern deep neural networks. One fundamental assumption is that convolutional kernels should be shared for all examples in a dataset. We propose conditionally parameterized…

Computer Vision and Pattern Recognition · Computer Science 2020-09-07 Brandon Yang , Gabriel Bender , Quoc V. Le , Jiquan Ngiam

In convolutional neural network-based character recognition, pooling layers play an important role in dimensionality reduction and deformation compensation. However, their kernel shapes and pooling operations are empirically predetermined;…

Computer Vision and Pattern Recognition · Computer Science 2021-07-13 Takato Otsuzuki , Heon Song , Seiichi Uchida , Hideaki Hayashi