SpecAugment is a very effective data augmentation method for both HMM and E2E-based automatic speech recognition (ASR) systems. Especially, it also works in low-resource scenarios. However, SpecAugment masks the spectrum of time or the frequency domain in a fixed augmentation policy, which may bring relatively less data diversity to the low-resource ASR. In this paper, we propose a policy-based SpecAugment (Policy-SpecAugment) method to alleviate the above problem. The idea is to use the augmentation-select policy and the augmentation-parameter changing policy to solve the fixed way. These policies are learned based on the loss of validation set, which is applied to the corresponding augmentation policies. It aims to encourage the model to learn more diverse data, which the model relatively requires. In experiments, we evaluate the effectiveness of our approach in low-resource scenarios, i.e., the 100 hours librispeech task. According to the results and analysis, we can see that the above issue can be obviously alleviated using our proposal. In addition, the experimental results show that, compared with the state-of-the-art SpecAugment, the proposed Policy-SpecAugment has a relative WER reduction of more than 10% on the Test/Dev-clean set, more than 5% on the Test/Dev-other set, and an absolute WER reduction of more than 1% on all test sets.
This paper analyzes the power imbalance issue in power-domain NOMA (PD-NOMA) in the presence of channel correlations, typically encountered on the downlink of cellular systems when the base station antennas have an insufficient separation. In a recent paper, the authors analyzed this issue for a typical uplink scenario with uncorrelated channels, and the study revealed an astounding result that the optimum in terms of average error probability is achieved when the user signals are perfectly balanced in terms of power as in multi-user MIMO with power control. This result led to some questioning of the concept of PD-NOMA for uncorrelated Rayleigh fading channels. In the present paper, we make a similar analysis for the downlink, and the study gives a very clear insight into the influence of the power imbalance at different levels of channel correlation. First, with full correlation (user signals transmitted from the same antenna), the PD-NOMA concept reduces to simple signal constellation design. The optimum is achieved when the power imbalance between the user signals is such that the resulting constellation has uniform spacing. Any deviation from this optimum will lead to a hierarchical constellation with performance loss. Also, this optimum power imbalance is shown to hold for a range of strong channel correlations, but for moderate and low correlation values perfectly power balanced NOMA takes over as in the presence of uncorrelated channels.
Intermediate layer output (ILO) regularization by means of multitask training on encoder side has been shown to be an effective approach to yielding improved results on a wide range of end-to-end ASR frameworks. In this paper, we propose a novel method to do ILO regularized training differently. Instead of using conventional multitask methods that entail more training overhead, we directly make the intermediate layer output as input to the decoder, that is, our decoder not only accepts the output of the final encoder layer as input, it also takes the output of the encoder ILO as input during training. With the proposed method, as both encoder and decoder are simultaneously "regularized", the network is more sufficiently trained, consistently leading to improved results, over the ILO-based CTC method, as well as over the original attention-based modeling method without the proposed method employed.
Internal Language Model Estimation (ILME) based language model (LM) fusion has been shown significantly improved recognition results over conventional shallow fusion in both intra-domain and cross-domain speech recognition tasks. In this paper, we attempt to apply our ILME method to cross-domain code-switching speech recognition (CSSR) work. Specifically, our curiosity comes from several aspects. First, we are curious about how effective the ILME-based LM fusion is for both intra-domain and cross-domain CSSR tasks. We verify this with or without merging two code-switching domains. More importantly, we train an end-to-end (E2E) speech recognition model by means of merging two monolingual data sets and observe the efficacy of the proposed ILME-based LM fusion for CSSR. Experimental results on SEAME that is from Southeast Asian and another Chinese Mainland CS data set demonstrate the effectiveness of the proposed ILME-based LM fusion method.
Speech emotion recognition (SER) is an essential part of human-computer interaction. In this paper, we propose an SER network based on a Graph Isomorphism Network with Weighted Multiple Aggregators (WMA-GIN), which can effectively handle the problem of information confusion when neighbour nodes' features are aggregated together in GIN structure. Moreover, a Full-Adjacent (FA) layer is adopted for alleviating the over-squashing problem, which is existed in all Graph Neural Network (GNN) structures, including GIN. Furthermore, a multi-phase attention mechanism and multi-loss training strategy are employed to avoid missing the useful emotional information in the stacked WMA-GIN layers. We evaluated the performance of our proposed WMA-GIN on the popular IEMOCAP dataset. The experimental results show that WMA-GIN outperforms other GNN-based methods and is comparable to some advanced non-graph-based methods by achieving 72.48% of weighted accuracy (WA) and 67.72% of unweighted accuracy (UA).
Sound event detection (SED) is an interesting but challenging task due to the scarcity of data and diverse sound events in real life. This paper presents a multi-grained based attention network (MGA-Net) for semi-supervised sound event detection. To obtain the feature representations related to sound events, a residual hybrid convolution (RH-Conv) block is designed to boost the vanilla convolution's ability to extract the time-frequency features. Moreover, a multi-grained attention (MGA) module is designed to learn temporal resolution features from coarse-level to fine-level. With the MGA module,the network could capture the characteristics of target events with short- or long-duration, resulting in more accurately determining the onset and offset of sound events. Furthermore, to effectively boost the performance of the Mean Teacher (MT) method, a spatial shift (SS) module as a data perturbation mechanism is introduced to increase the diversity of data. Experimental results show that the MGA-Net outperforms the published state-of-the-art competitors, achieving 53.27% and 56.96% event-based macro F1 (EB-F1) score, 0.709 and 0.739 polyphonic sound detection score (PSDS) on the validation and public set respectively.
This paper analyzes the power imbalance factor on the uplink of a 2-user Power-domain NOMA system and reveals that the minimum value of the average error probability is achieved when the user signals are perfectly balanced in terms of power as in Multi-User MIMO with power control. The analytic result is obtained by analyzing the pairwise error probability and exploiting a symmetry property of the error events. This result is supported by computer simulations using the QPSK and 16QAM signal formats and uncorrelated Rayleigh fading channels. This finding leads to the questioning of the basic philosophy of Power-domain NOMA and suggests that the best strategy for uncorrelated channels is to perfectly balance the average signal powers received from the users and to use a maximum likelihood receiver for their detection.
Low resource speech recognition has been long-suffering from insufficient training data. While neighbour languages are often used as assistant training data, it would be difficult for the model to induct similar units (character, subword, etc.) across the languages. In this paper, we assume similar units in neighbour language share similar term frequency and form a Huffman tree to perform multi-lingual hierarchical Softmax decoding. During decoding, the hierarchical structure can benefit the training of low-resource languages. Experimental results show the effectiveness of our method.
The performance of current Scene Graph Generation models is severely hampered by some hard-to-distinguish predicates, e.g., "woman-on/standing on/walking on-beach" or "woman-near/looking at/in front of-child". While general SGG models are prone to predict head predicates and existing re-balancing strategies prefer tail categories, none of them can appropriately handle these hard-to-distinguish predicates. To tackle this issue, inspired by fine-grained image classification, which focuses on differentiating among hard-to-distinguish object classes, we propose a method named Fine-Grained Predicates Learning (FGPL) which aims at differentiating among hard-to-distinguish predicates for Scene Graph Generation task. Specifically, we first introduce a Predicate Lattice that helps SGG models to figure out fine-grained predicate pairs. Then, utilizing the Predicate Lattice, we propose a Category Discriminating Loss and an Entity Discriminating Loss, which both contribute to distinguishing fine-grained predicates while maintaining learned discriminatory power over recognizable ones. The proposed model-agnostic strategy significantly boosts the performances of three benchmark models (Transformer, VCTree, and Motif) by 22.8\%, 24.1\% and 21.7\% of Mean Recall (mR@100) on the Predicate Classification sub-task, respectively. Our model also outperforms state-of-the-art methods by a large margin (i.e., 6.1\%, 4.6\%, and 3.2\% of Mean Recall (mR@100)) on the Visual Genome dataset.
In Uyghur speech, consonant and vowel reduction are often encountered, especially in spontaneous speech with high speech rate, which will cause a degradation of speech recognition performance. To solve this problem, we propose an effective phone mask training method for Conformer-based Uyghur end-to-end (E2E) speech recognition. The idea is to randomly mask off a certain percentage features of phones during model training, which simulates the above verbal phenomena and facilitates E2E model to learn more contextual information. According to experiments, the above issues can be greatly alleviated. In addition, deep investigations are carried out into different units in masking, which shows the effectiveness of our proposed masking unit. We also further study the masking method and optimize filling strategy of phone mask. Finally, compared with Conformer-based E2E baseline without mask training, our model demonstrates about 5.51% relative Word Error Rate (WER) reduction on reading speech and 12.92% on spontaneous speech, respectively. The above approach has also been verified on test-set of open-source data THUYG-20, which shows 20% relative improvements.