Video-to-speech synthesis is the task of reconstructing the speech signal from a silent video of a speaker. Most established approaches to date involve a two-step process, whereby an intermediate representation from the video, such as a spectrogram, is extracted first and then passed to a vocoder to produce the raw audio. Some recent work has focused on end-to-end synthesis, whereby the generation of raw audio and any intermediate representations is performed jointly. All such approaches involve training on data from almost exclusively audio-visual datasets, i.e. every audio sample has a corresponding video sample. This precludes the use of abundant audio-only datasets which may not have a corresponding visual modality (e.g. audiobooks, radio podcasts, speech recognition datasets etc.), as well as audio-only architectures that have been developed by the audio machine learning community over the years. In this paper we propose to train encoder-decoder models on more than 3,500 hours of audio data at 24kHz, and then use the pre-trained decoders to initialize the audio decoders for the video-to-speech synthesis task. The pre-training step uses audio samples only and does not require labels or corresponding samples from other modalities (visual, text). We demonstrate that this pre-training step improves the reconstructed speech and that it is an unexplored way to improve the quality of the generator in a cross-modal task while only requiring samples from one of the modalities. We conduct experiments using both raw audio and mel spectrograms as target outputs and benchmark our models with existing work.
Smartphone cameras today are increasingly approaching the versatility and quality of professional cameras through a combination of hardware and software advancements. However, fixed aperture remains a key limitation, preventing users from controlling the depth of field (DoF) of captured images. At the same time, many smartphones now have multiple cameras with different fixed apertures -- specifically, an ultra-wide camera with wider field of view and deeper DoF and a higher resolution primary camera with shallower DoF. In this work, we propose $\text{DC}^2$, a system for defocus control for synthetically varying camera aperture, focus distance and arbitrary defocus effects by fusing information from such a dual-camera system. Our key insight is to leverage real-world smartphone camera dataset by using image refocus as a proxy task for learning to control defocus. Quantitative and qualitative evaluations on real-world data demonstrate our system's efficacy where we outperform state-of-the-art on defocus deblurring, bokeh rendering, and image refocus. Finally, we demonstrate creative post-capture defocus control enabled by our method, including tilt-shift and content-based defocus effects.
Search engine plays a crucial role in satisfying users' diverse information needs. Recently, Pretrained Language Models (PLMs) based text ranking models have achieved huge success in web search. However, many state-of-the-art text ranking approaches only focus on core relevance while ignoring other dimensions that contribute to user satisfaction, e.g., document quality, recency, authority, etc. In this work, we focus on ranking user satisfaction rather than relevance in web search, and propose a PLM-based framework, namely SAT-Ranker, which comprehensively models different dimensions of user satisfaction in a unified manner. In particular, we leverage the capacities of PLMs on both textual and numerical inputs, and apply a multi-field input that modularizes each dimension of user satisfaction as an input field. Overall, SAT-Ranker is an effective, extensible, and data-centric framework that has huge potential for industrial applications. On rigorous offline and online experiments, SAT-Ranker obtains remarkable gains on various evaluation sets targeting different dimensions of user satisfaction. It is now fully deployed online to improve the usability of our search engine.
Table-to-text generation aims at automatically generating text to help people conveniently obtain salient information in tables. Recent works explicitly decompose the generation process into content planning and surface generation stages, employing two autoregressive networks for them respectively. However, they are computationally expensive due to the non-parallelizable nature of autoregressive decoding and the redundant parameters of two networks. In this paper, we propose the first totally non-autoregressive table-to-text model (Plan-then-Seam, PTS) that produces its outputs in parallel with one single network. PTS firstly writes and calibrates one plan of the content to be generated with a novel rethinking pointer predictor, and then takes the plan as the context for seaming to decode the description. These two steps share parameters and perform iteratively to capture token inter-dependency while keeping parallel decoding. Experiments on two public benchmarks show that PTS achieves 3.0~5.6 times speedup for inference time, reducing 50% parameters, while maintaining as least comparable performance against strong two-stage table-to-text competitors.
For text summarization, the role of discourse structure is pivotal in discerning the core content of a text. Regrettably, prior studies on incorporating Rhetorical Structure Theory (RST) into transformer-based summarization models only consider the nuclearity annotation, thereby overlooking the variety of discourse relation types. This paper introduces the 'RSTformer', a novel summarization model that comprehensively incorporates both the types and uncertainty of rhetorical relations. Our RST-attention mechanism, rooted in document-level rhetorical structure, is an extension of the recently devised Longformer framework. Through rigorous evaluation, the model proposed herein exhibits significant superiority over state-of-the-art models, as evidenced by its notable performance on several automatic metrics and human evaluation.
Constructing a highly accurate handwritten OCR system requires large amounts of representative training data, which is both time-consuming and expensive to collect. To mitigate the issue, we propose a denoising diffusion probabilistic model (DDPM) to generate training samples. This model conditions on a printed glyph image and creates mappings between printed characters and handwritten images, thus enabling the generation of photo-realistic handwritten samples with diverse styles and unseen text contents. However, the text contents in synthetic images are not always consistent with the glyph conditional images, leading to unreliable labels of synthetic samples. To address this issue, we further propose a progressive data filtering strategy to add those samples with a high confidence of correctness to the training set. Experimental results on IAM benchmark task show that OCR model trained with augmented DDPM-synthesized training samples can achieve about 45% relative word error rate reduction compared with the one trained on real data only.
We tackle the task of automatically discriminating between human and machine translations. As opposed to most previous work, we perform experiments in a multilingual setting, considering multiple languages and multilingual pretrained language models. We show that a classifier trained on parallel data with a single source language (in our case German-English) can still perform well on English translations that come from different source languages, even when the machine translations were produced by other systems than the one it was trained on. Additionally, we demonstrate that incorporating the source text in the input of a multilingual classifier improves (i) its accuracy and (ii) its robustness on cross-system evaluation, compared to a monolingual classifier. Furthermore, we find that using training data from multiple source languages (German, Russian, and Chinese) tends to improve the accuracy of both monolingual and multilingual classifiers. Finally, we show that bilingual classifiers and classifiers trained on multiple source languages benefit from being trained on longer text sequences, rather than on sentences.
The study of speech disorders can benefit greatly from time-aligned data. However, audio-text mismatches in disfluent speech cause rapid performance degradation for modern speech aligners, hindering the use of automatic approaches. In this work, we propose a simple and effective modification of alignment graph construction of CTC-based models using Weighted Finite State Transducers. The proposed weakly-supervised approach alleviates the need for verbatim transcription of speech disfluencies for forced alignment. During the graph construction, we allow the modeling of common speech disfluencies, i.e. repetitions and omissions. Further, we show that by assessing the degree of audio-text mismatch through the use of Oracle Error Rate, our method can be effectively used in the wild. Our evaluation on a corrupted version of the TIMIT test set and the UCLASS dataset shows significant improvements, particularly for recall, achieving a 23-25% relative improvement over our baselines.
Image manipulation under the guidance of textual descriptions has recently received a broad range of attention. In this study, we focus on the regional editing of images with the guidance of given text prompts. Different from current mask-based image editing methods, we propose a novel region-aware diffusion model (RDM) for entity-level image editing, which could automatically locate the region of interest and replace it following given text prompts. To strike a balance between image fidelity and inference speed, we design the intensive diffusion pipeline by combing latent space diffusion and enhanced directional guidance. In addition, to preserve image content in non-edited regions, we introduce regional-aware entity editing to modify the region of interest and preserve the out-of-interest region. We validate the proposed RDM beyond the baseline methods through extensive qualitative and quantitative experiments. The results show that RDM outperforms the previous approaches in terms of visual quality, overall harmonization, non-editing region content preservation, and text-image semantic consistency. The codes are available at https://github.com/haha-lisa/RDM-Region-Aware-Diffusion-Model.
While there has been significant development of models for Plain Language Summarization (PLS), evaluation remains a challenge. This is in part because PLS involves multiple, interrelated language transformations (e.g., adding background explanations, removing specialized terminology). No metrics are explicitly engineered for PLS, and the suitability of other text generation evaluation metrics remains unclear. To address these concerns, our study presents a granular meta-evaluation testbed, APPLS, designed to evaluate existing metrics for PLS. Drawing on insights from previous research, we define controlled perturbations for our testbed along four criteria that a metric of plain language should capture: informativeness, simplification, coherence, and faithfulness. Our analysis of metrics using this testbed reveals that current metrics fail to capture simplification, signaling a crucial gap. In response, we introduce POMME, a novel metric designed to assess text simplification in PLS. We demonstrate its correlation with simplification perturbations and validate across a variety of datasets. Our research contributes the first meta-evaluation testbed for PLS and a comprehensive evaluation of existing metrics, offering insights with relevance to other text generation tasks.