Zobrazeno 1 - 10
of 68
pro vyhledávání: '"Valpola, Harri"'
In many practical applications, large language models (LLMs) need to incorporate new knowledge not present in their pre-training data. The primary methods for this are fine-tuning and retrieval-augmented generation (RAG). Although RAG has emerged as
Externí odkaz:
http://arxiv.org/abs/2412.14964
Autor:
Boney, Rinu, Di Palo, Norman, Berglund, Mathias, Ilin, Alexander, Kannala, Juho, Rasmus, Antti, Valpola, Harri
Trajectory optimization using a learned model of the environment is one of the core elements of model-based reinforcement learning. This procedure often suffers from exploiting inaccuracies of the learned model. We propose to regularize trajectory op
Externí odkaz:
http://arxiv.org/abs/1903.11981
Autor:
Di Palo, Norman, Valpola, Harri
Model based predictions of future trajectories of a dynamical system often suffer from inaccuracies, forcing model based control algorithms to re-plan often, thus being computationally expensive, suboptimal and not reliable. In this work, we propose
Externí odkaz:
http://arxiv.org/abs/1812.03955
We prove an exact relationship between the optimal denoising function and the data distribution in the case of additive Gaussian noise, showing that denoising implicitly models the structure of data allowing it to be exploited in the unsupervised lea
Externí odkaz:
http://arxiv.org/abs/1709.02797
Autor:
Prémont-Schwarz, Isabeau, Ilin, Alexander, Hao, Tele Hotloo, Rasmus, Antti, Boney, Rinu, Valpola, Harri
We propose a recurrent extension of the Ladder networks whose structure is motivated by the inference required in hierarchical latent variable models. We demonstrate that the recurrent Ladder is able to handle a wide variety of complex learning tasks
Externí odkaz:
http://arxiv.org/abs/1707.09219
Autor:
Tarvainen, Antti, Valpola, Harri
The recently proposed Temporal Ensembling has achieved state-of-the-art results in several semi-supervised learning benchmarks. It maintains an exponential moving average of label predictions on each training example, and penalizes predictions that a
Externí odkaz:
http://arxiv.org/abs/1703.01780
Autor:
Greff, Klaus, Rasmus, Antti, Berglund, Mathias, Hao, Tele Hotloo, Schmidhuber, Jürgen, Valpola, Harri
We present a framework for efficient perceptual inference that explicitly reasons about the segmentation of its inputs and features. Rather than being trained for any specific segmentation, our framework learns the grouping process in an unsupervised
Externí odkaz:
http://arxiv.org/abs/1606.06724
We combine supervised learning with unsupervised learning in deep neural networks. The proposed model is trained to simultaneously minimize the sum of supervised and unsupervised cost functions by backpropagation, avoiding the need for layer-wise pre
Externí odkaz:
http://arxiv.org/abs/1507.02672
We show how a deep denoising autoencoder with lateral connections can be used as an auxiliary unsupervised learning task to support supervised learning. The proposed model is trained to minimize simultaneously the sum of supervised and unsupervised c
Externí odkaz:
http://arxiv.org/abs/1504.08215
Suitable lateral connections between encoder and decoder are shown to allow higher layers of a denoising autoencoder (dAE) to focus on invariant representations. In regular autoencoders, detailed information needs to be carried through the highest la
Externí odkaz:
http://arxiv.org/abs/1412.7210