Multilingual ASR offers training, deployment and overall performance benefits, but models trained via simple data pooling are known to suffer from cross-lingual interference. Oracle language information (exact-prior) and language-specific parameters are usually leveraged to overcome this, but such approaches cannot enable seamless, truly multilingual experiences. Existing methods try to overcome this limitation by relying on inferred language information or language agnostic mixture-of-experts, but they incur additional runtime complexity and/or training cost in addition to being less effective in streaming scenarios. Building on previous studies where models were trained to handle mixed-prior (knowledge that the underlying language belongs to a known group), we propose Attentive Multilingual Speech Encoding (AMuSE), a training framework designed to match exact-prior performance even in the absence of underlying language information at runtime (zero-prior), thereby making the model prior-agnostic. Leveraging AMuSE, we build a zero-prior enabled LLM-based ASR system that outperforms several exact-prior driven state-of-the-art benchmarks.

Contextualised word representations (e.g., ELMo and BERT) have been shown to outperform static representations (e.g., Word2vec, Fasttext, and GloVe) for many NLP tasks. In this paper, we investigate the use of contextualised embeddings for code search and classification, an area receiving less attention. We construct CodeELMo by training ELMo from scratch and fine tuning CodeBERT embeddings using masked language modeling based on natural language (NL) texts related to software development concepts and programming language (PL) texts consisting of method comment pairs from open source code bases. The dimensionality of the Finetuned Code BERT embeddings is reduced using linear transformations and augmented with a CodeELMo representation to develop CodeELBE – a lowdimensional contextualised software code representation. Results for binary classification and retrieval tasks show that CodeELBE1considerably improves retrieval performance on standard deep code search datasets compared to CodeBERT and baseline BERT models.

Learning to recognize actions from only a handful of labeled videos is a challenging problem due to the scarcity of tediously collected activity labels. We approach this problem by learning a two-pathway temporal contrastive model using unlabeled videos at two different speeds lever-aging the fact that changing video speed does not change an action. Specifically, we propose to maximize the similarity between encoded representations of the same video at two different speeds as well as minimize the similarity between different videos played at different speeds. This way we use the rich supervisory information in terms of ‘time’ that is present in otherwise unsupervised pool of videos. With this simple yet effective strategy of manipulating video playback rates, we considerably outperform video extensions of sophisticated state-of-the-art semi-supervised image recognition methods across multiple diverse bench-mark datasets and network architectures. Interestingly, our proposed approach benefits from out-of-domain unlabeled videos showing generalization and robustness. We also per-form rigorous ablations and analysis to validate our approach. Project page: https://cvir.github.io/TCL/.