Efficient Fine-tuning of Audio Spectrogram Transformers via Soft Mixture of Adapters

Mixture of Experts (MoE) architectures have recently started burgeoning due to their ability to scale model's capacity while maintaining the computational cost affordable. Furthermore, they can be applied to both Transformers and State Space Models, the current state-of-the-art models in numerous fields. While MoE has been mostly investigated for the pre-training stage, its use in parameter-efficient transfer learning settings is under-explored. To narrow this gap, this paper attempts to demystify the use of MoE for parameter-efficient fine-tuning of Audio Spectrogram Transformers to audio and speech downstream tasks. Specifically, we propose Soft Mixture of Adapters (Soft-MoA). It exploits adapters as the experts and, leveraging the recent Soft MoE method, it relies on a soft assignment between the input tokens and experts to keep the computational time limited. Extensive experiments across 4 benchmarks demonstrate that Soft-MoA outperforms the single adapter method and performs on par with the dense MoA counterpart. We finally present ablation studies on key elements of Soft-MoA, showing for example that Soft-MoA achieves better scaling with more experts, as well as ensuring that all experts contribute to the computation of the output tokens, thus dispensing with the expert imbalance issue.

Parameter-Efficient Transfer Learning of Audio Spectrogram Transformers

The common modus operandi of fine-tuning large pre-trained Transformer models entails the adaptation of all their parameters (i.e., full fine-tuning). While achieving striking results on multiple tasks, this approach becomes unfeasible as the model size and the number of downstream tasks increase. In natural language processing and computer vision, parameter-efficient approaches like prompt-tuning and adapters have emerged as solid alternatives by fine-tuning only a small number of extra parameters, without sacrificing performance accuracy. Specifically, adapters, due to their flexibility, have recently garnered significant attention, leading to several variants. For audio classification tasks, the Audio Spectrogram Transformer model shows impressive results. However, surprisingly, how to efficiently adapt it to several downstream tasks has not been tackled before. In this paper, we bridge this gap and present a detailed investigation of common parameter-efficient methods, revealing that adapters consistently outperform the other methods across four benchmarks. This trend is also confirmed in few-shot learning settings and when the total number of trainable parameters increases, demonstrating adapters superior scalability. We finally study the best adapter configuration, as well as the role of residual connections in the learning process.

Continual Contrastive Spoken Language Understanding

Recently, neural networks have shown impressive progress across diverse fields, with speech processing being no exception. However, recent breakthroughs in this area require extensive offline training using large datasets and tremendous computing resources. Unfortunately, these models struggle to retain their previously acquired knowledge when learning new tasks continually, and retraining from scratch is almost always impractical. In this paper, we investigate the problem of learning sequence-to-sequence models for spoken language understanding in a class-incremental learning (CIL) setting and we propose COCONUT, a CIL method that relies on the combination of experience replay and contrastive learning. Through a modified version of the standard supervised contrastive loss applied only to the rehearsal samples, COCONUT preserves the learned representations by pulling closer samples from the same class and pushing away the others. Moreover, we leverage a multimodal contrastive loss that helps the model learn more discriminative representations of the new data by aligning audio and text features. We also investigate different contrastive designs to combine the strengths of the contrastive loss with teacher-student architectures used for distillation. Experiments on two established SLU datasets reveal the effectiveness of our proposed approach and significant improvements over the baselines. We also show that COCONUT can be combined with methods that operate on the decoder side of the model, resulting in further metrics improvements.

Training dynamic models using early exits for automatic speech recognition on resource-constrained devices

The possibility of dynamically modifying the computational load of neural models at inference time is crucial for on-device processing, where computational power is limited and time-varying. Established approaches for neural model compression exist, but they provide architecturally static models. In this paper, we investigate the use of early-exit architectures, that rely on intermediate exit branches, applied to large-vocabulary speech recognition. This allows for the development of dynamic models that adjust their computational cost to the available resources and recognition performance. Unlike previous works, besides using pre-trained backbones we also train the model from scratch with an early-exit architecture. Experiments on public datasets show that early-exit architectures from scratch not only preserve performance levels when using fewer encoder layers, but also improve task accuracy as compared to using single-exit models or using pre-trained models. Additionally, we investigate an exit selection strategy based on posterior probabilities as an alternative to frame-based entropy.

An Experimental Review of Speaker Diarization methods with application to Two-Speaker Conversational Telephone Speech recordings

We performed an experimental review of current diarization systems for the conversational telephone speech (CTS) domain. In detail, we considered a total of eight different algorithms belonging to clustering-based, end-to-end neural diarization (EEND), and speech separation guided diarization (SSGD) paradigms. We studied the inference-time computational requirements and diarization accuracy on four CTS datasets with different characteristics and languages. We found that, among all methods considered, EEND-vector clustering (EEND-VC) offers the best trade-off in terms of computing requirements and performance. More in general, EEND models have been found to be lighter and faster in inference compared to clustering-based methods. However, they also require a large amount of diarization-oriented annotated data. In particular EEND-VC performance in our experiments degraded when the dataset size was reduced, whereas self-attentive EEND (SA-EEND) was less affected. We also found that SA-EEND gives less consistent results among all the datasets compared to EEND-VC, with its performance degrading on long conversations with high speech sparsity. Clustering-based diarization systems, and in particular VBx, instead have more consistent performance compared to SA-EEND but are outperformed by EEND-VC. The gap with respect to this latter is reduced when overlap-aware clustering methods are considered. SSGD is the most computationally demanding method, but it could be convenient if speech recognition has to be performed. Its performance is close to SA-EEND but degrades significantly when the training and inference data characteristics are less matched.

Sequence-Level Knowledge Distillation for Class-Incremental End-to-End Spoken Language Understanding

The ability to learn new concepts sequentially is a major weakness for modern neural networks, which hinders their use in non-stationary environments. Their propensity to fit the current data distribution to the detriment of the past acquired knowledge leads to the catastrophic forgetting issue. In this work we tackle the problem of Spoken Language Understanding applied to a continual learning setting. We first define a class-incremental scenario for the SLURP dataset. Then, we propose three knowledge distillation (KD) approaches to mitigate forgetting for a sequence-to-sequence transformer model: the first KD method is applied to the encoder output (audio-KD), and the other two work on the decoder output, either directly on the token-level (tok-KD) or on the sequence-level (seq-KD) distributions. We show that the seq-KD substantially improves all the performance metrics, and its combination with the audio-KD further decreases the average WER and enhances the entity prediction metric.

End-to-End Integration of Speech Separation and Voice Activity Detection for Low-Latency Diarization of Telephone Conversations

Recent works show that speech separation guided diarization (SSGD) is an increasingly promising direction, mainly thanks to the recent progress in speech separation. It performs diarization by first separating the speakers and then applying voice activity detection (VAD) on each separated stream. In this work we conduct an in-depth study of SSGD in the conversational telephone speech (CTS) domain, focusing mainly on low-latency streaming diarization applications. We consider three state-of-the-art speech separation (SSep) algorithms and study their performance both in online and offline scenarios, considering non-causal and causal implementations as well as continuous SSep (CSS) windowed inference. We compare different SSGD algorithms on two widely used CTS datasets: CALLHOME and Fisher Corpus (Part 1 and 2) and evaluate both separation and diarization performance. To improve performance, a novel, causal and computationally efficient leakage removal algorithm is proposed, which significantly decreases false alarms. We also explore, for the first time, fully end-to-end SSGD integration between SSep and VAD modules. Crucially, this enables fine-tuning on real-world data for which oracle speakers sources are not available. In particular, our best model achieves 8.8% DER on CALLHOME, which outperforms the current state-of-the-art end-to-end neural diarization model, despite being trained on an order of magnitude less data and having significantly lower latency, i.e., 0.1 vs. 1 seconds. Finally, we also show that the separated signals can be readily used also for automatic speech recognition, reaching performance close to using oracle sources in some configurations.

Improving the Intent Classification accuracy in Noisy Environment

Intent classification is a fundamental task in the spoken language understanding field that has recently gained the attention of the scientific community, mainly because of the feasibility of approaching it with end-to-end neural models. In this way, avoiding using intermediate steps, i.e. automatic speech recognition, is possible, thus the propagation of errors due to background noise, spontaneous speech, speaking styles of users, etc. Towards the development of solutions applicable in real scenarios, it is interesting to investigate how environmental noise and related noise reduction techniques to address the intent classification task with end-to-end neural models. In this paper, we experiment with a noisy version of the fluent speech command data set, combining the intent classifier with a time-domain speech enhancement solution based on Wave-U-Net and considering different training strategies. Experimental results reveal that, for this task, the use of speech enhancement greatly improves the classification accuracy in noisy conditions, in particular when the classification model is trained on enhanced signals.

Scaling strategies for on-device low-complexity source separation with Conv-Tasnet

Recently, several very effective neural approaches for single-channel speech separation have been presented in the literature. However, due to the size and complexity of these models, their use on low-resource devices, e.g. for hearing aids, and earphones, is still a challenge and established solutions are not available yet. Although approaches based on either pruning or compressing neural models have been proposed, the design of a model architecture suitable for a certain application domain often requires heuristic procedures not easily portable to different low-resource platforms. Given the modular nature of the well-known Conv-Tasnet speech separation architecture, in this paper we consider three parameters that directly control the overall size of the model, namely: the number of residual blocks, the number of repetitions of the separation blocks and the number of channels in the depth-wise convolutions, and experimentally evaluate how they affect the speech separation performance. In particular, experiments carried out on the Libri2Mix show that the number of dilated 1D-Conv blocks is the most critical parameter and that the usage of extra-dilation in the residual blocks allows reducing the performance drop.

Exploring the Joint Use of Rehearsal and Knowledge Distillation in Continual Learning for Spoken Language Understanding

Continual learning refers to a dynamical framework in which a model or agent receives a stream of non-stationary data over time and must adapt to new data while preserving previously acquired knowledge. Unfortunately, deep neural networks fail to meet these two desiderata, incurring the so-called catastrophic forgetting phenomenon. Whereas a vast array of strategies have been proposed to attenuate forgetting in the computer vision domain, for speech-related tasks, on the other hand, there is a dearth of works. In this paper, we turn our attention toward the joint use of rehearsal and knowledge distillation (KD) approaches for spoken language understanding under a class-incremental learning scenario. We report on multiple KD combinations at different levels in the network, showing that combining feature-level and predictions-level KDs leads to the best results. Finally, we provide an ablation study on the effect of the size of the rehearsal memory that corroborates the appropriateness of our approach for low-resource devices.