Alessandro Rozza
Research Subjects
Alessandro Rozza has a total of 22 co-authors. Here are the top co-authors by number of shared publications:
- Marco Leonardi
Lastminute.com Group, Vicolo de’ Calvi 2, 6830 Chiasso, Switzerland
5 shared publications - Paolo Napoletano
Department of Informatics, Systems and Communication, University of Milano-Bicocca, Viale Sarca 336, 20126 Milano, Italy
3 shared publications - A. Kasem
Lastminute.com Group, Vicolo de’ Calvi 2, 6830 Chiasso, Switzerland
2 shared publications - Adam Elwood
Aily Labs, Munich, Germany
2 shared publications - Lorenzo Ciampiconi
Lastminute.com Group, Vicolo de’ Calvi 2, 6830 Chiasso, Switzerland
2 shared publications - Mario Manzo
Information Technology Services, University of Naples “L’Orientale”, Naples, Italy
2 shared publications
Latest Publications
- Top Cited PapersOpen Access
A Survey and Taxonomy of Loss Functions in Machine Learning
Most state-of-the-art machine learning techniques revolve around the optimization of loss functions, making the choice of an objective critical to model performance and reliability. Although recent reviews discuss loss functions in specific domains or in deep learning settings, there is still no single reference that presents widely used losses across major task families within a unified formal setting and with consistent optimization-relevant property annotations. In this survey, we compile and systematize the most widely adopted loss functions for regression, classification, generative modeling, ranking, energy-based modeling, and relational learning. Our selection procedure combines seeding from foundational textbooks and prior surveys with cross-checking of highly cited literature and common implementations in mainstream machine learning frameworks. We introduce 52 loss functions and organize them into an intuitive taxonomy, summarizing their theoretical motivation, key mathematical properties, and typical application contexts, with compact appendix tables for quick lookup. This survey is intended as a resource for undergraduate, graduate, and Ph.D. students, as well as researchers seeking a structured reference for selecting and comparing loss functions.
Top Cited PapersOpen AccessA Survey and Taxonomy of Loss Functions in Machine Learning
Most state-of-the-art machine learning techniques revolve around the optimization of loss functions, making the choice of an objective critical to model performance and reliability. Although recent reviews discuss loss functions in specific domains or in deep learning settings, there is still no single reference that presents widely used losses across major task families within a unified formal setting and with consistent optimization-relevant property annotations. In this survey, we compile and systematize the most widely adopted loss functions for regression, classification, generative modeling, ranking, energy-based modeling, and relational learning. Our selection procedure combines seeding from foundational textbooks and prior surveys with cross-checking of highly cited literature and common implementations in mainstream machine learning frameworks. We introduce 52 loss functions and organize them into an intuitive taxonomy, summarizing their theoretical motivation, key mathematical properties, and typical application contexts, with compact appendix tables for quick lookup. This survey is intended as a resource for undergraduate, graduate, and Ph.D. students, as well as researchers seeking a structured reference for selecting and comparing loss functions.
- Open Access
A Primal-Dual Online Learning Approach for Dynamic Pricing of Sequentially Displayed Complementary Items under Sale Constraints
We address the challenging problem of dynamically pricing complementary items that are sequentially displayed to customers. An illustrative example is the online sale of flight tickets, where customers navigate through multiple web pages. Initially, they view the ticket cost, followed by ancillary expenses such as insurance and additional luggage fees. Coherent pricing policies for complementary items are essential because optimizing the pricing of each item individually is ineffective. Our scenario also involves a sales constraint, which specifies a minimum number of items to sell, and uncertainty regarding customer demand curves. To tackle this problem, we originally formulate it as a Markov Decision Process with constraints. Leveraging online learning tools, we design a primal-dual online optimization algorithm. We empirically evaluate our approach using synthetic settings randomly generated from real-world data, covering various configurations from stationary to non-stationary, and compare its performance in terms of constraints violation and regret against well-known baselines optimizing each state singularly.
Open AccessA Primal-Dual Online Learning Approach for Dynamic Pricing of Sequentially Displayed Complementary Items under Sale Constraints
We address the challenging problem of dynamically pricing complementary items that are sequentially displayed to customers. An illustrative example is the online sale of flight tickets, where customers navigate through multiple web pages. Initially, they view the ticket cost, followed by ancillary expenses such as insurance and additional luggage fees. Coherent pricing policies for complementary items are essential because optimizing the pricing of each item individually is ineffective. Our scenario also involves a sales constraint, which specifies a minimum number of items to sell, and uncertainty regarding customer demand curves. To tackle this problem, we originally formulate it as a Markov Decision Process with constraints. Leveraging online learning tools, we design a primal-dual online optimization algorithm. We empirically evaluate our approach using synthetic settings randomly generated from real-world data, covering various configurations from stationary to non-stationary, and compare its performance in terms of constraints violation and regret against well-known baselines optimizing each state singularly.
- 24 January 2024
- Published by Springer Nature
- p. 74-85
A General Purpose Method for Image Collection Summarization and Exploration
- 24 January 2024
- Published by Springer Nature
- p. 74-85
A General Purpose Method for Image Collection Summarization and Exploration
- Open Access
Maximum Entropy Exploration in Contextual Bandits with Neural Networks and Energy Based Models
Contextual bandits can solve a huge range of real-world problems. However, current popular algorithms to solve them either rely on linear models or unreliable uncertainty estimation in non-linear models, which are required to deal with the exploration–exploitation trade-off. Inspired by theories of human cognition, we introduce novel techniques that use maximum entropy exploration, relying on neural networks to find optimal policies in settings with both continuous and discrete action spaces. We present two classes of models, one with neural networks as reward estimators, and the other with energy based models, which model the probability of obtaining an optimal reward given an action. We evaluate the performance of these models in static and dynamic contextual bandit simulation environments. We show that both techniques outperform standard baseline algorithms, such as NN HMC, NN Discrete, Upper Confidence Bound, and Thompson Sampling, where energy based models have the best overall performance. This provides practitioners with new techniques that perform well in static and dynamic settings, and are particularly well suited to non-linear scenarios with continuous action spaces.
Open AccessMaximum Entropy Exploration in Contextual Bandits with Neural Networks and Energy Based Models
Contextual bandits can solve a huge range of real-world problems. However, current popular algorithms to solve them either rely on linear models or unreliable uncertainty estimation in non-linear models, which are required to deal with the exploration–exploitation trade-off. Inspired by theories of human cognition, we introduce novel techniques that use maximum entropy exploration, relying on neural networks to find optimal policies in settings with both continuous and discrete action spaces. We present two classes of models, one with neural networks as reward estimators, and the other with energy based models, which model the probability of obtaining an optimal reward given an action. We evaluate the performance of these models in static and dynamic contextual bandit simulation environments. We show that both techniques outperform standard baseline algorithms, such as NN HMC, NN Discrete, Upper Confidence Bound, and Thompson Sampling, where energy based models have the best overall performance. This provides practitioners with new techniques that perform well in static and dynamic settings, and are particularly well suited to non-linear scenarios with continuous action spaces.
- 22 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 31 (10577149) , 5009-5024
Composition and Style Attributes Guided Image Aesthetic Assessment
The aesthetic quality of an image is defined as the measure or appreciation of the beauty of an image. Aesthetics is inherently a subjective property but there are certain factors that influence it such as, the semantic content of the image, the attributes describing the artistic aspect, the photographic setup used for the shot, etc. In this paper we propose a method for the automatic prediction of the aesthetics of an image that is based on the analysis of the semantic content, the artistic style and the composition of the image. The proposed network includes: a pre-trained network for semantic features extraction (the Backbone); a Multi Layer Perceptron (MLP) network that relies on the Backbone features for the prediction of image attributes (the AttributeNet); a self-adaptive Hypernetwork that exploits the attributes prior encoded into the embedding generated by the AttributeNet to predict the parameters of the target network dedicated to aesthetic estimation (the AestheticNet). Given an image, the proposed multi-network is able to predict: style and composition attributes, and aesthetic score distribution. Results on three benchmark datasets demonstrate the effectiveness of the proposed method, while the ablation study gives a better understanding of the proposed network.
- 22 July 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Image Processing
- Vol. 31 (10577149) , 5009-5024
Composition and Style Attributes Guided Image Aesthetic Assessment
The aesthetic quality of an image is defined as the measure or appreciation of the beauty of an image. Aesthetics is inherently a subjective property but there are certain factors that influence it such as, the semantic content of the image, the attributes describing the artistic aspect, the photographic setup used for the shot, etc. In this paper we propose a method for the automatic prediction of the aesthetics of an image that is based on the analysis of the semantic content, the artistic style and the composition of the image. The proposed network includes: a pre-trained network for semantic features extraction (the Backbone); a Multi Layer Perceptron (MLP) network that relies on the Backbone features for the prediction of image attributes (the AttributeNet); a self-adaptive Hypernetwork that exploits the attributes prior encoded into the embedding generated by the AttributeNet to predict the parameters of the target network dedicated to aesthetic estimation (the AestheticNet). Given an image, the proposed multi-network is able to predict: style and composition attributes, and aesthetic score distribution. Results on three benchmark datasets demonstrate the effectiveness of the proposed method, while the ablation study gives a better understanding of the proposed network.
- 1 August 2021
- Published by Elsevier in Pattern Recognition Letters
- Vol. 148, 15-21
Graph-based neural network models with multiple self-supervised auxiliary tasks
- 1 August 2021
- Published by Elsevier in Pattern Recognition Letters
- Vol. 148, 15-21
Graph-based neural network models with multiple self-supervised auxiliary tasks
- Open Access
No Reference, Opinion Unaware Image Quality Assessment by Anomaly Detection
We propose an anomaly detection based image quality assessment method which exploits the correlations between feature maps from a pre-trained Convolutional Neural Network (CNN). The proposed method encodes the intra-layer correlation through the Gram matrix and then estimates the quality score combining the average of the correlation and the output from an anomaly detection method. The latter evaluates the degree of abnormality of an image by computing a correlation similarity with respect to a dictionary of pristine images. The effectiveness of the method is tested on different benchmarking datasets (LIVE-itW, KONIQ, and SPAQ).
Open AccessNo Reference, Opinion Unaware Image Quality Assessment by Anomaly Detection
We propose an anomaly detection based image quality assessment method which exploits the correlations between feature maps from a pre-trained Convolutional Neural Network (CNN). The proposed method encodes the intra-layer correlation through the Gram matrix and then estimates the quality score combining the average of the correlation and the output from an anomaly detection method. The latter evaluates the degree of abnormality of an image by computing a correlation similarity with respect to a dictionary of pristine images. The effectiveness of the method is tested on different benchmarking datasets (LIVE-itW, KONIQ, and SPAQ).
- 1 January 2020
- Published by Elsevier in Pattern Recognition
- Vol. 97
Dynamic graph convolutional networks
- 1 January 2020
- Published by Elsevier in Pattern Recognition
- Vol. 97
Dynamic graph convolutional networks
- Open Access
- 24 May 2019
- Published by MDPI AG in Machine Learning and Knowledge Extraction
- Vol. 1 (2) , 684-697
DOPSIE: Deep-Order Proximity and Structural Information Embedding
Graph-embedding algorithms map a graph into a vector space with the aim of preserving its structure and its intrinsic properties. Unfortunately, many of them are not able to encode the neighborhood information of the nodes well, especially from a topological prospective. To address this limitation, we propose a novel graph-embedding method called Deep-Order Proximity and Structural Information Embedding (DOPSIE). It provides topology and depth information at the same time through the analysis of the graph structure. Topological information is provided through clustering coefficients (CCs), which is connected to other structural properties, such as transitivity, density, characteristic path length, and efficiency, useful for representation in the vector space. The combination of individual node properties and neighborhood information constitutes an optimal network representation. Our experimental results show that DOPSIE outperforms state-of-the-art embedding methodologies in different classification problems.
- 24 May 2019
- Published by MDPI AG in Machine Learning and Knowledge Extraction
- Vol. 1 (2) , 684-697
Open AccessDOPSIE: Deep-Order Proximity and Structural Information Embedding
Graph-embedding algorithms map a graph into a vector space with the aim of preserving its structure and its intrinsic properties. Unfortunately, many of them are not able to encode the neighborhood information of the nodes well, especially from a topological prospective. To address this limitation, we propose a novel graph-embedding method called Deep-Order Proximity and Structural Information Embedding (DOPSIE). It provides topology and depth information at the same time through the analysis of the graph structure. Topological information is provided through clustering coefficients (CCs), which is connected to other structural properties, such as transitivity, density, characteristic path length, and efficiency, useful for representation in the vector space. The combination of individual node properties and neighborhood information constitutes an optimal network representation. Our experimental results show that DOPSIE outperforms state-of-the-art embedding methodologies in different classification problems.
- Open Access
Learning Combinations of Activation Functions
In the last decade, an active area of research has been devoted to design novel activation functions that are able to help deep neural networks to converge, obtaining better performance. The training procedure of these architectures usually involves optimization of the weights of their layers only, while non-linearities are generally pre-specified and their (possible) parameters are usually considered as hyper-parameters to be tuned manually. In this paper, we introduce two approaches to automatically learn different combinations of base activation functions (such as the identity function, ReLU, and tanh) during the training phase. We present a thorough comparison of our novel approaches with well-known architectures (such as LeNet-5, AlexNet, and ResNet-56) on three standard datasets (Fashion-MNIST, CIFAR-10, and ILSVRC-2012), showing substantial improvements in the overall performance, such as an increase in the top-1 accuracy for AlexNet on ILSVRC-2012 of 3.01 percentage points.
Open AccessLearning Combinations of Activation Functions
In the last decade, an active area of research has been devoted to design novel activation functions that are able to help deep neural networks to converge, obtaining better performance. The training procedure of these architectures usually involves optimization of the weights of their layers only, while non-linearities are generally pre-specified and their (possible) parameters are usually considered as hyper-parameters to be tuned manually. In this paper, we introduce two approaches to automatically learn different combinations of base activation functions (such as the identity function, ReLU, and tanh) during the training phase. We present a thorough comparison of our novel approaches with well-known architectures (such as LeNet-5, AlexNet, and ResNet-56) on three standard datasets (Fashion-MNIST, CIFAR-10, and ILSVRC-2012), showing substantial improvements in the overall performance, such as an increase in the top-1 accuracy for AlexNet on ILSVRC-2012 of 3.01 percentage points.