The Latest Papers about AI
-
Direct simultaneous speech to speech translation
2021-10-15 17:59:15Xutai Ma, Hongyu Gong, Danni Liu, Ann Lee, Yun Tang, Peng-Jen Chen, Wei-Ning Hsu, Kenneth Heafield, Phillip Koehn, Juan PinoAbstract
We present the first direct simultaneous speech-to-speech translation (Simul-S2ST) model, with the ability to start generating translation in the target speech before consuming the full source speech content and independently from intermediate text representations. Our approach leverages recent progress on direct speech-to-speech translation with discrete units. Instead of continuous spectrogram features, a sequence of direct representations, which are learned in a unsupervised manner, are predicted from the model and passed directly to a vocoder for speech synthesis. The simultaneous policy then operates on source speech features and target discrete units. Finally, a vocoder synthesize the target speech from discrete units on-the-fly. We carry out numerical studies to compare cascaded and direct approach on Fisher Spanish-English dataset.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08250
PDF
-
Textual Backdoor Attacks Can Be More Harmful via Two Simple Tricks
2021-10-15 17:58:46Yangyi Chen, Fanchao Qi, Zhiyuan Liu, Maosong SunAbstract
Backdoor attacks are a kind of emergent security threat in deep learning. When a deep neural model is injected with a backdoor, it will behave normally on standard inputs but give adversary-specified predictions once the input contains specific backdoor triggers. Current textual backdoor attacks have poor attack performance in some tough situations. In this paper, we find two simple tricks that can make existing textual backdoor attacks much more harmful. The first trick is to add an extra training task to distinguish poisoned and clean data during the training of the victim model, and the second one is to use all the clean training data rather than remove the original clean data corresponding to the poisoned data. These two tricks are universally applicable to different attack models. We conduct experiments in three tough situations including clean data fine-tuning, low poisoning rate, and label-consistent attacks. Experimental results show that the two tricks can significantly improve attack performance. This paper exhibits the great potential harmfulness of backdoor attacks. All the code and data will be made public to facilitate further research.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08247
PDF
-
Tricks for Training Sparse Translation Models
2021-10-15 17:58:45Dheeru Dua, Shruti Bhosale, Vedanuj Goswami, James Cross, Mike Lewis, Angela FanAbstract
Multi-task learning with an unbalanced data distribution skews model learning towards high resource tasks, especially when model capacity is fixed and fully shared across all tasks. Sparse scaling architectures, such as BASELayers, provide flexible mechanisms for different tasks to have a variable number of parameters, which can be useful to counterbalance skewed data distributions. We find that that sparse architectures for multilingual machine translation can perform poorly out of the box, and propose two straightforward techniques to mitigate this - a temperature heating mechanism and dense pre-training. Overall, these methods improve performance on two multilingual translation benchmarks compared to standard BASELayers and Dense scaling baselines, and in combination, more than 2x model convergence speed.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08246
PDF
-
LPRules: Rule Induction in Knowledge Graphs Using Linear Programming
2021-10-15 17:58:16Sanjeeb Dash, Joao GoncalvesAbstract
Knowledge graph (KG) completion is a well-studied problem in AI. Rule-based methods and embedding-based methods form two of the solution techniques. Rule-based methods learn first-order logic rules that capture existing facts in an input graph and then use these rules for reasoning about missing facts. A major drawback of such methods is the lack of scalability to large datasets. In this paper, we present a simple linear programming (LP) model to choose rules from a list of candidate rules and assign weights to them. For smaller KGs, we use simple heuristics to create the candidate list. For larger KGs, we start with a small initial candidate list, and then use standard column generation ideas to add more rules in order to improve the LP model objective value. To foster interpretability and generalizability, we limit the complexity of the set of chosen rules via explicit constraints, and tune the complexity hyperparameter for individual datasets. We show that our method can obtain state-of-the-art results for three out of four widely used KG datasets, while taking significantly less computing time than other popular rule learners including some based on neuro-symbolic methods. The improved scalability of our method allows us to tackle large datasets such as YAGO3-10.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08245
PDF
-
Performance, Successes and Limitations of Deep Learning Semantic Segmentation of Multiple Defects in Transmission Electron Micrographs
2021-10-15 17:57:59Ryan Jacobs, Mingren Shen, Yuhan Liu, Wei Hao, Xiaoshan Li, Ruoyu He, Jacob RC Greaves, Donglin Wang, Zeming Xie, Zitong Huang, Chao Wang, Kevin G. Field, Dane MorganAbstract
In this work, we perform semantic segmentation of multiple defect types in electron microscopy images of irradiated FeCrAl alloys using a deep learning Mask Regional Convolutional Neural Network (Mask R-CNN) model. We conduct an in-depth analysis of key model performance statistics, with a focus on quantities such as predicted distributions of defect shapes, defect sizes, and defect areal densities relevant to informing modeling and understanding of irradiated Fe-based materials properties. To better understand the performance and present limitations of the model, we provide examples of useful evaluation tests which include a suite of random splits, and dataset size-dependent and domain-targeted cross validation tests. Overall, we find that the current model is a fast, effective tool for automatically characterizing and quantifying multiple defect types in microscopy images, with a level of accuracy on par with human domain expert labelers. More specifically, the model can achieve average defect identification F1 scores as high as 0.8, and, based on random cross validation, have low overall average (+/- standard deviation) defect size and density percentage errors of 7.3 (+/- 3.8)% and 12.7 (+/- 5.3)%, respectively. Further, our model predicts the expected material hardening to within 10-20 MPa (about 10% of total hardening), which is about the same error level as experiments. Our targeted evaluation tests also suggest the best path toward improving future models is not expanding existing databases with more labeled images but instead data additions that target weak points of the model domain, such as images from different microscopes, imaging conditions, irradiation environments, and alloy types. Finally, we discuss the first phase of an effort to provide an easy-to-use, open-source object detection tool to the broader community for identifying defects in new images.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08244
PDF
-
Neural Dubber: Dubbing for Silent Videos According to Scripts
2021-10-15 17:56:07Chenxu Hu, Qiao Tian, Tingle Li, Yuping Wang, Yuxuan Wang, Hang ZhaoAbstract
Dubbing is a post-production process of re-recording actors' dialogues, which is extensively used in filmmaking and video production. It is usually performed manually by professional voice actors who read lines with proper prosody, and in synchronization with the pre-recorded videos. In this work, we propose Neural Dubber, the first neural network model to solve a novel automatic video dubbing (AVD) task: synthesizing human speech synchronized with the given silent video from the text. Neural Dubber is a multi-modal text-to-speech (TTS) model that utilizes the lip movement in the video to control the prosody of the generated speech. Furthermore, an image-based speaker embedding (ISE) module is developed for the multi-speaker setting, which enables Neural Dubber to generate speech with a reasonable timbre according to the speaker's face. Experiments on the chemistry lecture single-speaker dataset and LRS2 multi-speaker dataset show that Neural Dubber can generate speech audios on par with state-of-the-art TTS models in terms of speech quality. Most importantly, both qualitative and quantitative evaluations show that Neural Dubber can control the prosody of synthesized speech by the video, and generate high-fidelity speech temporally synchronized with the video.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08243
PDF
-
Intent-based Product Collections for E-commerce using Pretrained Language Models
2021-10-15 17:52:42Hiun Kim, Jisu Jeong, Kyung-Min Kim, Dongjun Lee, Hyun Dong Lee, Dongpil Seo, Jeeseung Han, Dong Wook Park, Ji Ae Heo, Rak Yeong KimAbstract
Building a shopping product collection has been primarily a human job. With the manual efforts of craftsmanship, experts collect related but diverse products with common shopping intent that are effective when displayed together, e.g., backpacks, laptop bags, and messenger bags for freshman bag gifts. Automatically constructing a collection requires an ML system to learn a complex relationship between the customer's intent and the product's attributes. However, there have been challenging points, such as 1) long and complicated intent sentences, 2) rich and diverse product attributes, and 3) a huge semantic gap between them, making the problem difficult. In this paper, we use a pretrained language model (PLM) that leverages textual attributes of web-scale products to make intent-based product collections. Specifically, we train a BERT with triplet loss by setting an intent sentence to an anchor and corresponding products to positive examples. Also, we improve the performance of the model by search-based negative sampling and category-wise positive pair augmentation. Our model significantly outperforms the search-based baseline model for intent-based product matching in offline evaluations. Furthermore, online experimental results on our e-commerce platform show that the PLM-based method can construct collections of products with increased CTR, CVR, and order-diversity compared to expert-crafted collections.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08241
PDF
-
Fire Together Wire Together: A Dynamic Pruning Approach with Self-Supervised Mask Prediction
2021-10-15 17:39:53Sara Elkerdawy, Mostafa Elhoushi, Hong Zhang, Nilanjan RayAbstract
Dynamic model pruning is a recent direction that allows for the inference of a different sub-network for each input sample during deployment. However, current dynamic methods rely on learning a continuous channel gating through regularization by inducing sparsity loss. This formulation introduces complexity in balancing different losses (e.g task loss, regularization loss). In addition, regularization-based methods lack transparent tradeoff hyperparameter selection to realize computational budget. Our contribution is twofold: 1) decoupled task and pruning training. 2) Simple hyperparameter selection that enables FLOPs reduction estimation before training. We propose to predict a mask to process k filters in a layer based on the activation of its previous layer. We pose the problem as a self-supervised binary classification problem. Each mask predictor module is trained to predict if the log-likelihood of each filter in the current layer belongs to the top-k activated filters. The value k is dynamically estimated for each input based on a novel criterion using the mass of heatmaps. We show experiments on several neural architectures, such as VGG, ResNet, and MobileNet on CIFAR and ImageNet datasets. On CIFAR, we reach similar accuracy to SOTA methods with 15% and 24% higher FLOPs reduction. Similarly in ImageNet, we achieve a lower drop in accuracy with up to 13% improvement in FLOPs reduction.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08232
PDF
-
Cross-Domain Data Integration for Named Entity Disambiguation in Biomedical Text
2021-10-15 17:38:16Maya Varma, Laurel Orr, Sen Wu, Megan Leszczynski, Xiao Ling, Christopher RéAbstract
Named entity disambiguation (NED), which involves mapping textual mentions to structured entities, is particularly challenging in the medical domain due to the presence of rare entities. Existing approaches are limited by the presence of coarse-grained structural resources in biomedical knowledge bases as well as the use of training datasets that provide low coverage over uncommon resources. In this work, we address these issues by proposing a cross-domain data integration method that transfers structural knowledge from a general text knowledge base to the medical domain. We utilize our integration scheme to augment structural resources and generate a large biomedical NED dataset for pretraining. Our pretrained model with injected structural knowledge achieves state-of-the-art performance on two benchmark medical NED datasets: MedMentions and BC5CDR. Furthermore, we improve disambiguation of rare entities by up to 57 accuracy points.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08228
PDF
-
Guiding Visual Question Generation
2021-10-15 17:38:08Nihir Vedd, Zixu Wang, Marek Rei, Yishu Miao, Lucia SpeciaAbstract
In traditional Visual Question Generation (VQG), most images have multiple concepts (e.g. objects and categories) for which a question could be generated, but models are trained to mimic an arbitrary choice of concept as given in their training data. This makes training difficult and also poses issues for evaluation -- multiple valid questions exist for most images but only one or a few are captured by the human references. We present Guiding Visual Question Generation - a variant of VQG which conditions the question generator on categorical information based on expectations on the type of question and the objects it should explore. We propose two variants: (i) an explicitly guided model that enables an actor (human or automated) to select which objects and categories to generate a question for; and (ii) an implicitly guided model that learns which objects and categories to condition on, based on discrete latent variables. The proposed models are evaluated on an answer-category augmented VQA dataset and our quantitative results show a substantial improvement over the current state of the art (over 9 BLEU-4 increase). Human evaluation validates that guidance helps the generation of questions that are grammatically coherent and relevant to the given image and objects.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08226
PDF
-
DialFact: A Benchmark for Fact-Checking in Dialogue
2021-10-15 17:34:35Prakhar Gupta, Chien-Sheng Wu, Wenhao Liu, Caiming XiongAbstract
Fact-checking is an essential tool to mitigate the spread of misinformation and disinformation, however, it has been often explored to verify formal single-sentence claims instead of casual conversational claims. To study the problem, we introduce the task of fact-checking in dialogue. We construct DialFact, a testing benchmark dataset of 22,245 annotated conversational claims, paired with pieces of evidence from Wikipedia. There are three sub-tasks in DialFact: 1) Verifiable claim detection task distinguishes whether a response carries verifiable factual information; 2) Evidence retrieval task retrieves the most relevant Wikipedia snippets as evidence; 3) Claim verification task predicts a dialogue response to be supported, refuted, or not enough information. We found that existing fact-checking models trained on non-dialogue data like FEVER fail to perform well on our task, and thus, we propose a simple yet data-efficient solution to effectively improve fact-checking performance in dialogue. We point out unique challenges in DialFact such as handling the colloquialisms, coreferences, and retrieval ambiguities in the error analysis to shed light on future research in this direction.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08222
PDF
-
Combining Diverse Feature Priors
Abstract
To improve model generalization, model designers often restrict the features that their models use, either implicitly or explicitly. In this work, we explore the design space of leveraging such feature priors by viewing them as distinct perspectives on the data. Specifically, we find that models trained with diverse sets of feature priors have less overlapping failure modes, and can thus be combined more effectively. Moreover, we demonstrate that jointly training such models on additional (unlabeled) data allows them to correct each other's mistakes, which, in turn, leads to better generalization and resilience to spurious correlations. Code available at this https URL.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08220
PDF
-
Incremental Speech Synthesis For Speech-To-Speech Translation
2021-10-15 17:20:28Danni Liu, Changhan Wang, Hongyu Gong, Xutai Ma, Yun Tang, Juan PinoAbstract
In a speech-to-speech translation (S2ST) pipeline, the text-to-speech (TTS) module is an important component for delivering the translated speech to users. To enable incremental S2ST, the TTS module must be capable of synthesizing and playing utterances while its input text is still streaming in. In this work, we focus on improving the incremental synthesis performance of TTS models. With a simple data augmentation strategy based on prefixes, we are able to improve the incremental TTS quality to approach offline performance. Furthermore, we bring our incremental TTS system to the practical scenario in combination with an upstream simultaneous speech translation system, and show the gains also carry over to this use-case. In addition, we propose latency metrics tailored to S2ST applications, and investigate methods for latency reduction in this context.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08214
PDF
-
Towards Identity Preserving Normal to Dysarthric Voice Conversion
2021-10-15 17:18:02Wen-Chin Huang, Bence Mark Halpern, Lester Phillip Violeta, Odette Scharenborg, Tomoki TodaAbstract
We present a voice conversion framework that converts normal speech into dysarthric speech while preserving the speaker identity. Such a framework is essential for (1) clinical decision making processes and alleviation of patient stress, (2) data augmentation for dysarthric speech recognition. This is an especially challenging task since the converted samples should capture the severity of dysarthric speech while being highly natural and possessing the speaker identity of the normal speaker. To this end, we adopted a two-stage framework, which consists of a sequence-to-sequence model and a nonparallel frame-wise model. Objective and subjective evaluations were conducted on the UASpeech dataset, and results showed that the method was able to yield reasonable naturalness and capture severity aspects of the pathological speech. On the other hand, the similarity to the normal source speaker's voice was limited and requires further improvements.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08213
PDF
-
Multitask Prompted Training Enables Zero-Shot Task Generalization
2021-10-15 17:08:57Victor Sanh, Albert Webson, Colin Raffel, Stephen H. Bach, Lintang Sutawika, Zaid Alyafeai, Antoine Chaffin, Arnaud Stiegler, Teven Le Scao, Arun Raja, Manan Dey, M Saiful Bari, Canwen Xu, Urmish Thakker, Shanya Sharma Sharma, Eliza Szczechla, Taewoon Kim, Gunjan Chhablani, Nihal Nayak, Debajyoti Datta, Jonathan Chang, Mike Tian-Jian Jiang, Han Wang, Matteo Manica, Sheng Shen, Zheng Xin Yong, Harshit Pandey, Rachel Bawden, Thomas Wang, Trishala Neeraj, Jos Rozen, Abheesht Sharma, Andrea Santilli, Thibault Fevry, Jason Alan Fries, Ryan Teehan, Stella Biderman, Leo Gao, Tali Bers, Thomas Wolf, Alexander M. RushAbstract
Large language models have recently been shown to attain reasonable zero-shot generalization on a diverse set of tasks. It has been hypothesized that this is a consequence of implicit multitask learning in language model training. Can zero-shot generalization instead be directly induced by explicit multitask learning? To test this question at scale, we develop a system for easily mapping general natural language tasks into a human-readable prompted form. We convert a large set of supervised datasets, each with multiple prompts using varying natural language. These prompted datasets allow for benchmarking the ability of a model to perform completely unseen tasks specified in natural language. We fine-tune a pretrained encoder-decoder model on this multitask mixture covering a wide variety of tasks. The model attains strong zero-shot performance on several standard datasets, often outperforming models 16x its size. Further, our approach attains strong performance on a subset of tasks from the BIG-Bench benchmark, outperforming models 6x its size. All prompts and trained models are available at this http URL.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08207
PDF
-
Shared Visual Representations of Drawing for Communication: How do different biases affect human interpretability and intent?
Abstract
We present an investigation into how representational losses can affect the drawings produced by artificial agents playing a communication game. Building upon recent advances, we show that a combination of powerful pretrained encoder networks, with appropriate inductive biases, can lead to agents that draw recognisable sketches, whilst still communicating well. Further, we start to develop an approach to help automatically analyse the semantic content being conveyed by a sketch and demonstrate that current approaches to inducing perceptual biases lead to a notion of objectness being a key feature despite the agent training being self-supervised.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08203
PDF
-
Evaluation of Hyperparameter-Optimization Approaches in an Industrial Federated Learning System
2021-10-15 17:01:40Stephanie Holly, Thomas Hiessl, Safoura Rezapour Lakani, Daniel Schall, Clemens Heitzinger, Jana KemnitzAbstract
Federated Learning (FL) decouples model training from the need for direct access to the data and allows organizations to collaborate with industry partners to reach a satisfying level of performance without sharing vulnerable business information. The performance of a machine learning algorithm is highly sensitive to the choice of its hyperparameters. In an FL setting, hyperparameter optimization poses new challenges. In this work, we investigated the impact of different hyperparameter optimization approaches in an FL system. In an effort to reduce communication costs, a critical bottleneck in FL, we investigated a local hyperparameter optimization approach that -- in contrast to a global hyperparameter optimization approach -- allows every client to have its own hyperparameter configuration. We implemented these approaches based on grid search and Bayesian optimization and evaluated the algorithms on the MNIST data set using an i.i.d. partition and on an Internet of Things (IoT) sensor based industrial data set using a non-i.i.d. partition.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08202
PDF
-
BBQ: A Hand-Built Bias Benchmark for Question Answering
2021-10-15 16:43:46Alicia Parrish, Angelica Chen, Nikita Nangia, Vishakh Padmakumar, Jason Phang, Jana Thompson, Phu Mon Htut, Samuel R. BowmanAbstract
It is well documented that NLP models learn social biases present in the world, but little work has been done to show how these biases manifest in actual model outputs for applied tasks like question answering (QA). We introduce the Bias Benchmark for QA (BBQ), a dataset consisting of question-sets constructed by the authors that highlight \textit{attested} social biases against people belonging to protected classes along nine different social dimensions relevant for U.S. English-speaking contexts. Our task evaluates model responses at two distinct levels: (i) given an under-informative context, test how strongly model answers reflect social biases, and (ii) given an adequately informative context, test whether the model's biases still override a correct answer choice. We find that models strongly rely on stereotypes when the context is ambiguous, meaning that the model's outputs consistently reproduce harmful biases in this setting. Though models are much more accurate when the context provides an unambiguous answer, they still rely on stereotyped information and achieve an accuracy 2.5 percentage points higher on examples where the correct answer aligns with a social bias, with this accuracy difference widening to 5 points for examples targeting gender.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08193
PDF
-
Why don't people use character-level machine translation?
2021-10-15 16:43:31Jindřich Libovický, Helmut Schmid, Alexander FraserAbstract
We present a literature and empirical survey that critically assesses the state of the art in character-level modeling for machine translation (MT). Despite evidence in the literature that character-level systems are comparable with subword systems, they are virtually never used in competitive setups in WMT competitions. We empirically show that even with recent modeling innovations in character-level natural language processing, character-level MT systems still struggle to match their subword-based counterparts both in terms of translation quality and training and inference speed. Character-level MT systems show neither better domain robustness, nor better morphological generalization, despite being often so motivated. On the other hand, they tend to be more robust towards source side noise and the translation quality does not degrade with increasing beam size at decoding time.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08191
PDF
-
Attention meets Geometry: Geometry Guided Spatial-Temporal Attention for Consistent Self-Supervised Monocular Depth Estimation
2021-10-15 16:43:31Patrick Ruhkamp, Daoyi Gao, Hanzhi Chen, Nassir Navab, Benjamin BusamAbstract
Inferring geometrically consistent dense 3D scenes across a tuple of temporally consecutive images remains challenging for self-supervised monocular depth prediction pipelines. This paper explores how the increasingly popular transformer architecture, together with novel regularized loss formulations, can improve depth consistency while preserving accuracy. We propose a spatial attention module that correlates coarse depth predictions to aggregate local geometric information. A novel temporal attention mechanism further processes the local geometric information in a global context across consecutive images. Additionally, we introduce geometric constraints between frames regularized by photometric cycle consistency. By combining our proposed regularization and the novel spatial-temporal-attention module we fully leverage both the geometric and appearance-based consistency across monocular frames. This yields geometrically meaningful attention and improves temporal depth stability and accuracy compared to previous methods.
Abstract (translated)
URL
https://arxiv.org/abs/2110.08192
PDF
