Zobrazeno 1 - 10
of 162
pro vyhledávání: '"Trancoso, Isabel"'
Speech is a rich biomarker that encodes substantial information about the health of a speaker, and thus it has been proposed for the detection of numerous diseases, achieving promising results. However, questions remain about what the models trained
Externí odkaz:
http://arxiv.org/abs/2409.10230
Although human evaluation remains the gold standard for open-domain dialogue evaluation, the growing popularity of automated evaluation using Large Language Models (LLMs) has also extended to dialogue. However, most frameworks leverage benchmarks tha
Externí odkaz:
http://arxiv.org/abs/2408.10902
Despite being heralded as the new standard for dialogue evaluation, the closed-source nature of GPT-4 poses challenges for the community. Motivated by the need for lightweight, open source, and multilingual dialogue evaluators, this paper introduces
Externí odkaz:
http://arxiv.org/abs/2407.11660
Large Language Models (LLMs) have showcased remarkable capabilities in various Natural Language Processing tasks. For automatic open-domain dialogue evaluation in particular, LLMs have been seamlessly integrated into evaluation frameworks, and togeth
Externí odkaz:
http://arxiv.org/abs/2407.03841
Autor:
Teixeira, Francisco, Pizzi, Karla, Olivier, Raphael, Abad, Alberto, Raj, Bhiksha, Trancoso, Isabel
Membership Inference (MI) poses a substantial privacy threat to the training data of Automatic Speech Recognition (ASR) systems, while also offering an opportunity to audit these models with regard to user data. This paper explores the effectiveness
Externí odkaz:
http://arxiv.org/abs/2405.01207
Autor:
Mendonça, John, Pereira, Patrícia, Menezes, Miguel, Cabarrão, Vera, Farinha, Ana C., Moniz, Helena, Carvalho, João Paulo, Lavie, Alon, Trancoso, Isabel
Task-oriented conversational datasets often lack topic variability and linguistic diversity. However, with the advent of Large Language Models (LLMs) pretrained on extensive, multilingual and diverse text data, these limitations seem overcome. Nevert
Externí odkaz:
http://arxiv.org/abs/2311.13910
Publikováno v:
IEEE Access, vol. 12, pp. 82949-82971, 2024
Speaker embeddings are ubiquitous, with applications ranging from speaker recognition and diarization to speech synthesis and voice anonymisation. The amount of information held by these embeddings lends them versatility, but also raises privacy conc
Externí odkaz:
http://arxiv.org/abs/2310.06652
Autor:
Mendonça, John, Pereira, Patrícia, Moniz, Helena, Carvalho, João Paulo, Lavie, Alon, Trancoso, Isabel
Despite significant research effort in the development of automatic dialogue evaluation metrics, little thought is given to evaluating dialogues other than in English. At the same time, ensuring metrics are invariant to semantically similar responses
Externí odkaz:
http://arxiv.org/abs/2308.16797
The main limiting factor in the development of robust multilingual dialogue evaluation metrics is the lack of multilingual data and the limited availability of open sourced multilingual dialogue systems. In this work, we propose a workaround for this
Externí odkaz:
http://arxiv.org/abs/2308.16795
Automatic Speaker Diarization (ASD) is an enabling technology with numerous applications, which deals with recordings of multiple speakers, raising special concerns in terms of privacy. In fact, in remote settings, where recordings are shared with a
Externí odkaz:
http://arxiv.org/abs/2210.14995