Interleaved spoken language models (SLMs) alternately generate text and speech tokens, but decoding at full transformer depth for every step becomes costly, especially due to long speech sequences. We propose SPAR-K, a modality-aware early exit framework designed to accelerate interleaved SLM inference while preserving perceptual quality. SPAR-K introduces a speech alternating-depth schedule: most speech positions exit at a fixed intermediate layer, while periodic full-depth "refresh" steps mitigate distribution shift due to early exit. We evaluate our framework using Step-Audio-2-mini and GLM-4-Voice across four datasets spanning reasoning, factual QA, and dialogue tasks, measuring performance in terms of ASR transcription accuracy and perceptual quality. Experimental results demonstrate that SPAR-K largely preserves question-answering accuracy with a maximum accuracy drop of 0.82\% while reducing average speech decoding depth by up to 11\% on Step-Audio-2-mini and 5\% on GLM-4-Voice, both with negligible changes in MOS and WER and no auxiliary computation overhead. We further demonstrate that confidence-based early exit strategies, widely used in text LLMs, are suboptimal for SLMs, highlighting that the unique statistical nature of speech tokens necessitates a specialized early exit design.
Magahi is an Indo-Aryan Language, spoken mainly in the Eastern parts of India. Despite having a significant number of speakers, there has been virtually no language resource (LR) or language technology (LT) developed for the language, mainly because of its status as a non-scheduled language. The present paper describes an attempt to develop an annotated corpus of Magahi. The data is mainly taken from a couple of blogs in Magahi, some collection of stories in Magahi and the recordings of conversation in Magahi and it is annotated at the POS level using BIS tagset.
We investigate the efficiency of two very different spoken term detection approaches for transcription when the available data is insufficient to train a robust ASR system. This work is grounded in very low-resource language documentation scenario where only few minutes of recording have been transcribed for a given language so far.Experiments on two oral languages show that a pretrained universal phone recognizer, fine-tuned with only a few minutes of target language speech, can be used for spoken term detection with a better overall performance than a dynamic time warping approach. In additi…
Dr. Bhimrao Ambedkar University（宾希姆拉奥·安贝德卡尔博士大学）；Indian Institute of Technology - Kharagpur（印度理工学院克勒格布尔分校）
In the present paper, we will present a survey of the language resources and technologies available for the non-scheduled and endangered languages of India. While there have been different estimates from different sources about the number of languages in India, it could be assumed that there are more than 1,000 languages currently being spoken in India. However barring some of the 22 languages included in the 8th Schedule of the Indian Constitution (called the scheduled languages), there is hardly any substantial resource or technology available for the rest of the languages. Nonetheless there…
End-to-end spoken language understanding (SLU) systems benefit from pretraining on large corpora, followed by fine-tuning on application-specific data. The resulting models are too large for on-edge applications. For instance, BERT-based systems contain over 110M parameters. Observing the model is overparameterized, we propose lean transformer structure where the dimension of the attention mechanism is automatically reduced using group sparsity. We propose a variant where the learned attention subspace is transferred to an attention bottleneck layer. In a low-resource setting and without pre-t…
Evaluating a Multi-Agent Voice-Enabled Smart Speaker for Care Homes: A Safety-Focused Framework
University of Hull, UK（赫尔大学）；University of Southampton, UK（南安普顿大学）；Connexin, Hull, UK（Connexin公司）；Leeds Teaching Hospital NHS Trust, UK（利兹教学医院国家健康服务信托）
IndicTrans2: Towards High-Quality and Accessible Machine Translation Models for all 22 Scheduled Indian Languages
Nilekani Centre at AI4Bharat（AI4Bharat 尼勒卡尼中心）；Indian Institute of Technology Madras（马德拉斯印度理工学院）；Microsoft（微软）；National Institute of Information and Communications Technology（信息通信研究机构）；Flipkart（弗利普卡特）
Spoken Language Understanding (SLU) converts user utterances into structured semantic representations. Data sparsity is one of the main obstacles of SLU due to the high cost of human annotation, especially when domain changes or a new domain comes. In this work, we propose a data augmentation method with atomic templates for SLU, which involves minimum human efforts. The atomic templates produce exemplars for fine-grained constituents of semantic representations. We propose an encoder-decoder model to generate the whole utterance from atomic exemplars. Moreover, the generator could be transfer…
This paper presents the design of the machine learning architecture that underlies the Alexa Skills Kit (ASK) a large scale Spoken Language Understanding (SLU) Software Development Kit (SDK) that enables developers to extend the capabilities of Amazon's virtual assistant, Alexa. At Amazon, the infrastructure powers over 25,000 skills deployed through the ASK, as well as AWS's Amazon Lex SLU Service. The ASK emphasizes flexibility, predictability and a rapid iteration cycle for third party developers. It imposes inductive biases that allow it to learn robust SLU models from extremely small and …
Ping An Technology (Shenzhen) Co., Ltd.（平安科技（深圳）有限公司）
End-to-end Spoken Language Understanding (SLU) models are made increasingly large and complex to achieve the state-ofthe-art accuracy. However, the increased complexity of a model can also introduce high risk of over-fitting, which is a major challenge in SLU tasks due to the limitation of available data. In this paper, we propose an attention-based SLU model together with three encoder enhancement strategies to overcome data sparsity challenge. The first strategy focuses on the transferlearning approach to improve feature extraction capability of the encoder. It is implemented by pre-training…
Dementia is a growing problem as our society ages, and detection methods are often invasive and expensive. Recent deep-learning techniques can offer a faster diagnosis and have shown promising results. However, they require large amounts of labelled data which is not easily available for the task of dementia detection. One effective solution to sparse data problems is data augmentation, though the exact methods need to be selected carefully. To date, there has been no empirical study of data augmentation on Alzheimer's disease (AD) datasets for NLP and speech processing. In this work, we inves…
The University of Melbourne（墨尔本大学）；The University of Queensland（昆士兰大学）
Although the Indonesian language is spoken by almost 200 million people and the 10th most spoken language in the world, it is under-represented in NLP research. Previous work on Indonesian has been hampered by a lack of annotated datasets, a sparsity of language resources, and a lack of resource standardization. In this work, we release the IndoLEM dataset comprising seven tasks for the Indonesian language, spanning morpho-syntax, semantics, and discourse. We additionally release IndoBERT, a new pre-trained language model for Indonesian, and evaluate it over IndoLEM, in addition to benchmarkin…
Part of Speech (POS) tagging is crucial to Natural Language Processing (NLP). It is a well-studied topic in several resource-rich languages. However, the development of computational linguistic resources is still in its infancy despite the existence of numerous languages that are historically and literary rich. Assamese, an Indian scheduled language, spoken by more than 25 million people, falls under this category. In this paper, we present a Deep Learning (DL)-based POS tagger for Assamese. The development process is divided into two stages. In the first phase, several pre-trained word embedd…
BLSP: Bootstrapping Language-Speech Pre-training via Behavior Alignment of Continuation Writing
Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；School of Artificial Intelligence, University of Chinese Academy of Sciences（中国科学院大学人工智能学院）；Alibaba Inc.（阿里巴巴集团）；Machine Intelligence Technology Lab, Alibaba Inc.（阿里巴巴集团机器智能技术实验室）；Wuhan AI Research（武汉人工智能研究院）
