Export Citations
Publications
Save this search
Please login to be able to save your searches and receive alerts for new content matching your search criteria.
- research-article
Parameter prediction for unseen deep architectures
NIPS '21: Proceedings of the 35th International Conference on Neural Information Processing Systems•December 2021, Article No.: 2253, pp 29433-29448Deep learning has been successful in automating the design of features in machine learning pipelines. However, the algorithms optimizing neural network parameters remain largely hand-designed and computationally inefficient. We study if we can use deep ...
- research-article
Brick-by-brick: combinatorial construction with deep reinforcement learning
NIPS '21: Proceedings of the 35th International Conference on Neural Information Processing Systems•December 2021, Article No.: 439, pp 5745-5757Discovering a solution in a combinatorial space is prevalent in many real-world problems but it is also challenging due to diverse complex constraints and the vast number of possible combinations. To address such a problem, we introduce a novel ...
- research-article
Model zoos: a dataset of diverse populations of neural network models
NIPS '22: Proceedings of the 36th International Conference on Neural Information Processing Systems•November 2022, Article No.: 2763, pp 38134-38148In the last years, neural networks (NN) have evolved from laboratory environments to the state-of-the-art for many real-world problems. It was shown that NN models (i.e., their weights and biases) evolve on unique trajectories in weight space during ...
- research-article
Hyper-representations as generative models: sampling unseen neural network weights
NIPS '22: Proceedings of the 36th International Conference on Neural Information Processing Systems•November 2022, Article No.: 2024, pp 27906-27920Learning representations of neural network weights given a model zoo is an emerging and challenging area with many potential applications from model inspection, to neural architecture search or knowledge distillation. Recently, an autoencoder trained on ...
- research-article
Can we scale transformers to predict parameters of diverse imagenet models?
ICML'23: Proceedings of the 40th International Conference on Machine Learning•July 2023, Article No.: 709, pp 17243-17259Pretraining a neural network on a large dataset is becoming a cornerstone in machine learning that is within the reach of only a few communities with large-resources. We aim at an ambitious goal of democratizing pretraining. Towards that goal, we train ...
- research-article
Understanding attention and generalization in graph neural networks
NIPS'19: Proceedings of the 33rd International Conference on Neural Information Processing Systems•December 2019, Article No.: 378, pp 4202-4212We aim to better understand attention over nodes in graph neural networks (GNNs) and identify factors influencing its effectiveness. We particularly focus on the ability of attention GNNs to generalize to larger, more complex or noisy graphs. Motivated by ...
- research-article
Leveraging Large Face Recognition Data for Emotion Classification
2018 13th IEEE International Conference on Automatic Face & Gesture Recognition (FG 2018)•May 2018, pp 692-696• https://doi.org/10.1109/FG.2018.00109In this paper we describe a solution to our entry for the emotion recognition challenge EmotiW 2017. We propose an ensemble of several models, which capture spatial and audio features from videos. Spatial features are captured by convolutional neural ...
- research-article
Human nonverbal behavior multi-sourced ontological annotation
VIGTA '13: Proceedings of the International Workshop on Video and Image Ground Truth in Computer Vision Applications•July 2013, Article No.: 2, pp 1-8• https://doi.org/10.1145/2501105.2501111In this paper we introduce the current results of an ongoing three-year research and development project on automatic annotation of human nonverbal behavior. The present output of the project is a tool that provides algorithms and graphical user ...