Text-to-speech synthesis (TTS) is a task to convert texts into speech. Two of the factors that have been driving TTS are the advancements of probabilistic models and latent representation learning. We propose a TTS method based on latent variable conversion using a diffusion probabilistic model and the variational autoencoder (VAE). In our TTS method, we use a waveform model based on VAE, a diffusion model that predicts the distribution of latent variables in the waveform model from texts, and an alignment model that learns alignments between the text and speech latent sequences. Our method integrates diffusion with VAE by modeling both mean and variance parameters with diffusion, where the target distribution is determined by approximation from VAE. This latent variable conversion framework potentially enables us to flexibly incorporate various latent feature extractors. Our experiments show that our method is robust to linguistic labels with poor orthography and alignment errors.

End-to-end text-to-speech synthesis (TTS) can generate highly natural synthetic speech from raw text. However, rendering the correct pitch accents is still a challenging problem for end-to-end TTS. To tackle the challenge of rendering correct pitch accent in Japanese end-to-end TTS, we adopt PnG~BERT, a self-supervised pretrained model in the character and phoneme domain for TTS. We investigate the effects of features captured by PnG~BERT on Japanese TTS by modifying the fine-tuning condition to determine the conditions helpful inferring pitch accents. We manipulate content of PnG~BERT features from being text-oriented to speech-oriented by changing the number of fine-tuned layers during TTS. In addition, we teach PnG~BERT pitch accent information by fine-tuning with tone prediction as an additional downstream task. Our experimental results show that the features of PnG~BERT captured by pretraining contain information helpful inferring pitch accent, and PnG~BERT outperforms baseline Tacotron on accent correctness in a listening test.

Agents that can follow language instructions are expected to be useful in a variety of situations such as navigation. However, training neural network-based agents requires numerous paired trajectories and languages. This paper proposes using multimodal generative models for semi-supervised learning in the instruction following tasks. The models learn a shared representation of the paired data, and enable semi-supervised learning by reconstructing unpaired data through the representation. Key challenges in applying the models to sequence-to-sequence tasks including instruction following are learning a shared representation of variable-length mulitimodal data and incorporating attention mechanisms. To address the problems, this paper proposes a novel network architecture to absorb the difference in the sequence lengths of the multimodal data. In addition, to further improve the performance, this paper shows how to incorporate the generative model-based approach with an existing semi-supervised method called a speaker-follower model, and proposes a regularization term that improves inference using unpaired trajectories. Experiments on BabyAI and Room-to-Room (R2R) environments show that the proposed method improves the performance of instruction following by leveraging unpaired data, and improves the performance of the speaker-follower model by 2\% to 4\% in R2R.

A polarization camera has great potential for 3D reconstruction since the angle of polarization (AoP) and the degree of polarization (DoP) of reflected light are related to an object's surface normal. In this paper, we propose a novel 3D reconstruction method called Polarimetric Multi-View Inverse Rendering (Polarimetric MVIR) that effectively exploits geometric, photometric, and polarimetric cues extracted from input multi-view color-polarization images. We first estimate camera poses and an initial 3D model by geometric reconstruction with a standard structure-from-motion and multi-view stereo pipeline. We then refine the initial model by optimizing photometric rendering errors and polarimetric errors using multi-view RGB, AoP, and DoP images, where we propose a novel polarimetric cost function that enables an effective constraint on the estimated surface normal of each vertex, while considering four possible ambiguous azimuth angles revealed from the AoP measurement. The weight for the polarimetric cost is effectively determined based on the DoP measurement, which is regarded as the reliability of polarimetric information. Experimental results using both synthetic and real data demonstrate that our Polarimetric MVIR can reconstruct a detailed 3D shape without assuming a specific surface material and lighting condition.

Edema is a common symptom of kidney disease, and quantitative measurement of edema is desired. This paper presents a method to estimate the degree of edema from facial images taken before and after dialysis of renal failure patients. As tasks to estimate the degree of edema, we perform pre- and post-dialysis classification and body weight prediction. We develop a multi-patient pre-training framework for acquiring knowledge of edema and transfer the pre-trained model to a model for each patient. For effective pre-training, we propose a novel contrastive representation learning, called weight-aware supervised momentum contrast (WeightSupMoCo). WeightSupMoCo aims to make feature representations of facial images closer in similarity of patient weight when the pre- and post-dialysis labels are the same. Experimental results show that our pre-training approach improves the accuracy of pre- and post-dialysis classification by 15.1% and reduces the mean absolute error of weight prediction by 0.243 kg compared with training from scratch. The proposed method accurately estimate the degree of edema from facial images; our edema estimation system could thus be beneficial to dialysis patients.

Peripheral blood oxygen saturation (SpO2), an indicator of oxygen levels in the blood, is one of the most important physiological parameters. Although SpO2 is usually measured using a pulse oximeter, non-contact SpO2 estimation methods from facial or hand videos have been attracting attention in recent years. In this paper, we propose an SpO2 estimation method from facial videos based on convolutional neural networks (CNN). Our method constructs CNN models that consider the direct current (DC) and alternating current (AC) components extracted from the RGB signals of facial videos, which are important in the principle of SpO2 estimation. Specifically, we extract the DC and AC components from the spatio-temporal map using filtering processes and train CNN models to predict SpO2 from these components. We also propose an end-to-end model that predicts SpO2 directly from the spatio-temporal map by extracting the DC and AC components via convolutional layers. Experiments using facial videos and SpO2 data from 50 subjects demonstrate that the proposed method achieves a better estimation performance than current state-of-the-art SpO2 estimation methods.

Off-policy evaluation (OPE) attempts to predict the performance of counterfactual policies using log data from a different policy. We extend its applicability by developing an OPE method for a class of both full support and deficient support logging policies in contextual-bandit settings. This class includes deterministic bandit (such as Upper Confidence Bound) as well as deterministic decision-making based on supervised and unsupervised learning. We prove that our method's prediction converges in probability to the true performance of a counterfactual policy as the sample size increases. We validate our method with experiments on partly and entirely deterministic logging policies. Finally, we apply it to evaluate coupon targeting policies by a major online platform and show how to improve the existing policy.

We propose a novel backpropagation algorithm for training spiking neural networks (SNNs) that encodes information in the relative multiple spike timing of individual neurons without single-spike restrictions. The proposed algorithm inherits the advantages of conventional timing-based methods in that it computes accurate gradients with respect to spike timing, which promotes ideal temporal coding. Unlike conventional methods where each neuron fires at most once, the proposed algorithm allows each neuron to fire multiple times. This extension naturally improves the computational capacity of SNNs. Our SNN model outperformed comparable SNN models and achieved as high accuracy as non-convolutional artificial neural networks. The spike count property of our networks was altered depending on the time constant of the postsynaptic current and the membrane potential. Moreover, we found that there existed the optimal time constant with the maximum test accuracy. That was not seen in conventional SNNs with single-spike restrictions on time-to-fast-spike (TTFS) coding. This result demonstrates the computational properties of SNNs that biologically encode information into the multi-spike timing of individual neurons. Our code would be publicly available.

This paper presents a portrait stylization method designed for real-time mobile applications with limited style examples available. Previous learning based stylization methods suffer from the geometric and semantic gaps between portrait domain and style domain, which obstacles the style information to be correctly transferred to the portrait images, leading to poor stylization quality. Based on the geometric prior of human facial attributions, we propose to utilize geometric alignment to tackle this issue. Firstly, we apply Thin-Plate-Spline (TPS) on feature maps in the generator network and also directly to style images in pixel space, generating aligned portrait-style image pairs with identical landmarks, which closes the geometric gaps between two domains. Secondly, adversarial learning maps the textures and colors of portrait images to the style domain. Finally, geometric aware cycle consistency preserves the content and identity information unchanged, and deformation invariant constraint suppresses artifacts and distortions. Qualitative and quantitative comparison validate our method outperforms existing methods, and experiments proof our method could be trained with limited style examples (100 or less) in real-time (more than 40 FPS) on mobile devices. Ablation study demonstrates the effectiveness of each component in the framework.

Sequence transducers, such as the RNN-T and the Conformer-T, are one of the most promising models of end-to-end speech recognition, especially in streaming scenarios where both latency and accuracy are important. Although various methods, such as alignment-restricted training and FastEmit, have been studied to reduce the latency, latency reduction is often accompanied with a significant degradation in accuracy. We argue that this suboptimal performance might be caused because none of the prior methods explicitly model and reduce the latency. In this paper, we propose a new training method to explicitly model and reduce the latency of sequence transducer models. First, we define the expected latency at each diagonal line on the lattice, and show that its gradient can be computed efficiently within the forward-backward algorithm. Then we augment the transducer loss with this expected latency, so that an optimal trade-off between latency and accuracy is achieved. Experimental results on the WSJ dataset show that the proposed minimum latency training reduces the latency of causal Conformer-T from 220 ms to 27 ms within a WER degradation of 0.7%, and outperforms conventional alignment-restricted training (110 ms) and FastEmit (67 ms) methods.