Recently, RNN-Transducers have achieved remarkable results on various automatic speech recognition tasks. However, lattice-free sequence discriminative training methods, which obtain superior performance in hybrid modes, are rarely investigated in RNN-Transducers. In this work, we propose three lattice-free training objectives, namely lattice-free maximum mutual information, lattice-free segment-level minimum Bayes risk, and lattice-free minimum Bayes risk, which are used for the final posterior output of the phoneme-based neural transducer with a limited context dependency. Compared to criteria using N-best lists, lattice-free methods eliminate the decoding step for hypotheses generation during training, which leads to more efficient training. Experimental results show that lattice-free methods gain up to 6.5% relative improvement in word error rate compared to a sequence-level cross-entropy trained model. Compared to the N-best-list based minimum Bayes risk objectives, lattice-free methods gain 40% - 70% relative training time speedup with a small degradation in performance.

ASR can be improved by multi-task learning (MTL) with domain enhancing or domain adversarial training, which are two opposite objectives with the aim to increase/decrease domain variance towards domain-aware/agnostic ASR, respectively. In this work, we study how to best apply these two opposite objectives with speaker labels to improve conformer-based ASR. We also propose a novel adaptive gradient reversal layer for stable and effective adversarial training without tuning effort. Detailed analysis and experimental verification are conducted to show the optimal positions in the ASR neural network (NN) to apply speaker enhancing and adversarial training. We also explore their combination for further improvement, achieving the same performance as i-vectors plus adversarial training. Our best speaker-based MTL achieves 7\% relative improvement on the Switchboard Hub5'00 set. We also investigate the effect of such speaker-based MTL w.r.t. cleaner dataset and weaker ASR NN.

The pre-training of masked language models (MLMs) consumes massive computation to achieve good results on downstream NLP tasks, resulting in a large carbon footprint. In the vanilla MLM, the virtual tokens, [MASK]s, act as placeholders and gather the contextualized information from unmasked tokens to restore the corrupted information. It raises the question of whether we can append [MASK]s at a later layer, to reduce the sequence length for earlier layers and make the pre-training more efficient. We show: (1) [MASK]s can indeed be appended at a later layer, being disentangled from the word embedding; (2) The gathering of contextualized information from unmasked tokens can be conducted with a few layers. By further increasing the masking rate from 15% to 50%, we can pre-train RoBERTa-base and RoBERTa-large from scratch with only 78% and 68% of the original computational budget without any degradation on the GLUE benchmark. When pre-training with the original budget, our method outperforms RoBERTa for 6 out of 8 GLUE tasks, on average by 0.4%.

Automatic speech recognition (ASR) has been established as a well-performing technique for many scenarios where lots of labeled data is available. Additionally, unsupervised representation learning recently helped to tackle tasks with limited data. Following this, hardware limitations and applications give rise to the question how to efficiently take advantage of large pretrained models and reduce their complexity for downstream tasks. In this work, we study a challenging low resource conversational telephony speech corpus from the medical domain in Vietnamese and German. We show the benefits of using unsupervised techniques beyond simple fine-tuning of large pre-trained models, discuss how to adapt them to a practical telephony task including bandwidth transfer and investigate different data conditions for pre-training and fine-tuning. We outperform the project baselines by 22% relative using pretraining techniques. Further gains of 29% can be achieved by refinements of architecture and training and 6% by adding 0.8 h of in-domain adaptation data.

演讲者的适应性对于建立强大的自动语音识别（ASR）系统很重要。在这项工作中，我们根据基于配置符号的声学模型（AM）在300H数据集中的功能空间方法研究了扬声器自适应训练（SAT）的各种方法。我们提出了一种称为加权简单添加的方法，该方法将加权的说话者信息向量添加到构象异构体AM的多头自发动模块的输入中。使用此方法用于SAT，我们在HUB5'00和HUB5'01的Callhome部分方面取得了3.5％和4.5％的相对改善。此外，我们以先前的作品为基础，在此基础上，我们为基于构象异构体的混合动力AM提出了一种新颖的竞争培训配方。我们扩展并改善了此食谱，在该配方中，我们在打电筒300H HUB5'00数据集上的单词误差（WER）方面取得了11％的相对改善。我们还通过将参数总数减少34％，从而使该配方有效。

作为语音识别的最流行的序列建模方法之一，RNN-Transducer通过越来越复杂的神经网络模型，以增长的规模和增加训练时代的增长，实现了不断发展的性能。尽管强大的计算资源似乎是培训卓越模型的先决条件，但我们试图通过仔细设计更有效的培训管道来克服它。在这项工作中，我们提出了一条高效的三阶段渐进式训练管道，以在合理的短时间内从头开始建立具有非常有限的计算资源的高效神经传感器模型。每个阶段的有效性在LibrisPeech和Convebobly Corpora上都经过实验验证。拟议的管道能够在短短2-3周内以单个GPU接近最先进的性能来训练换能器模型。我们最好的构型传感器在Librispeech测试中获得4.1％的速度，仅使用35个训练时代。

本文总结了我们提交第10次对话系统技术挑战（DSTC10）的第二次赛道的任务2的任务2“关于口语对话的知识接地的任务导向对话建模”。类似于前一年的迭代，任务由三个子任务组成：检测转弯是否是知识寻求，选择相关知识文档并最终生成接地响应。今年，焦点在于将系统适应嘈杂的ASR成绩单。我们探讨了不同的方法，使模型对这种类型的输入更加强大，并使生成的响应适应口语对话的风格。对于后者，我们通过嘈杂的频道模型获得最佳效果，该模型另外减少了短和通用响应的数量。我们最好的系统在挑战的人类评估中实现了第一级和第三名。

为了减轻必须在神经语言模型的SoftMax归一化中遍历全面词汇的问题，在大型词汇基神经语言模型的背景下提出并研究了基于样本的培训标准。这些培训标准通常享有更快的培训和测试的好处，以困惑的困惑性能略微降低，几乎没有单词错误率下降。虽然噪声对比估计是最受欢迎的选择之一，但最近我们表明，只要完成额外的校正步骤，即可从原始模型输出中恢复预期的类后概率，而其他基于样本的基于基于的标准也可以表现良好。在这项工作中，我们提出了自我规范化的重要性抽样。与我们以前的工作相比，在这项工作中考虑的标准是自我规范化的，并且没有必要进一步进行更正步骤。与噪声对比估计相比，我们的方法在应用中的复杂性方面直接相当。通过自我规范化的语言模型培训以及格子救援实验，我们展示了我们提出的自我规范化重要性采样在面向研究的和以生产为导向的自动语音识别任务中具有竞争力。

最近提出的符合者架构已成功用于实现在不同数据集上实现最先进性能的端到端自动语音识别（ASR）架构。为了我们的最佳知识，没有研究使用适用物声学模型对混合ASR的影响。在本文中，我们展示并评估了竞争的基于统一体的混合模型训练配方。我们研究了不同的培训方面和方法，以提高字差率以及提高训练速度。我们应用时间下采样方法以实现有效的培训，并使用转换卷积再次上置输出序列。我们在交换机300H数据集中进行实验，与其他架构相比，我们的符合子的混合模型实现了竞争力。它在Hub5'01测试集上概括并显着优于BLSTM的混合模型。

Modeling perception sensors is key for simulation based testing of automated driving functions. Beyond weather conditions themselves, sensors are also subjected to object dependent environmental influences like tire spray caused by vehicles moving on wet pavement. In this work, a novel modeling approach for spray in lidar data is introduced. The model conforms to the Open Simulation Interface (OSI) standard and is based on the formation of detection clusters within a spray plume. The detections are rendered with a simple custom ray casting algorithm without the need of a fluid dynamics simulation or physics engine. The model is subsequently used to generate training data for object detection algorithms. It is shown that the model helps to improve detection in real-world spray scenarios significantly. Furthermore, a systematic real-world data set is recorded and published for analysis, model calibration and validation of spray effects in active perception sensors. Experiments are conducted on a test track by driving over artificially watered pavement with varying vehicle speeds, vehicle types and levels of pavement wetness. All models and data of this work are available open source.