Hover over the dots to explore related posts. Closer dots are more semantically related, and the red dot marks the current page.
Abstract¶
We focus on improving the effectiveness of a Virtual Assistant (VA) in recognizing emerging entities in spoken queries. We introduce a method that uses historical user interactions to forecast which entities will gain in popularity and become trending, and it subse- quently integrates the predictions within the Automated Speech Recognition (ASR) component of the VA. Experiments show that our proposed approach results in a 20% relative reduction in errors on emerging entity name utterances without degrading the overall recognition quality of the system.
Happy to share the news about my first joint pubication with the Siri Speech team at Apple. Our short paper Predicting Entity Popularity to Improve Spoken Entity Recognition by Virtual Assistants with Christophe van Gysel, myself, Ernie Pusateri, and Ilya Oparin, is accepted at SIGIR 2020.
Related Posts
-
Happy to share yet another publication with the Siri Speech team at Apple, this time led by Sashank Gondala, who interned with us last year. Our full paperError-driven Pruning of Language Models for Virtual Assistants is accepted at ICASSP 2021.
-
Abstract
User generated spoken audio remains a challenge for Automatic Speech Recognition (ASR) technology and content-based audio surrogates derived from ASR-transcripts must be error robust. An investigation of the use of term clouds as surrogates for podcasts demonstrates that ASR term clouds closely approximate term clouds derived from human-generated transcripts across a range of cloud sizes. A user study confirms the conclusion that ASR-clouds are viable surrogates for depicting the content of podcasts.
References
[1] Manos Tsagkias, Martha Larson, and Maarten de Rijke. 2008. Term clouds as surrogates for user generated speech. In Proceedings of the 31st annual international ACM SIGIR conference on Research and development in information retrieval (SIGIR ‘08). Association for Computing Machinery, New York, NY, USA, 773–774. ACM Link PDF
-
I took part in SIREN 2008, a research event in the Netherlands, presenting our work with Martha Larson and Maarten de Rijke, on Term Clouds as Surrogates for User Generated Speech [1]; see post
References
[1] Manos Tsagkias, Martha Larson, and Maarten de Rijke. 2008. Term clouds as surrogates for user generated speech. In Proceedings of the 31st annual international ACM SIGIR conference on Research and development in information retrieval (SIGIR ‘08). Association for Computing Machinery, New York, NY, USA, 773–774. ACM Link PDF
-
Figure 1. At the stage of AI2Future.
AI2Future is a local, Croatian, initiative for disseminating the importance and for grounding the use of Artificial Intelligence. The conference brings together researchers, A.I. start ups, and large corporates from all around Croatia to share experiences and learn from each other on what A.I. can and cannot do.
I was invited to give one of the two keynote talks. The first keynote focused on conversational agents and natural language processing and mine followed up with insights how search powers a large spectrum of applications from search, question and answering, and recommendations. I focused on the work we do at 904Labs and illustrated the principles of online learning to rank through real-world examples from our experimences with our customers.
-
Abstract
This paper describes the University of Amsterdam’s participation in the short track of the Entity Recognition & Disambiguation Challenge 2014 (ERD 2014). We describe how we adapt the Semanticizer—an open-source entity linking framework developed primarily at the University of Amsterdam—to the task of the ERD challenge: linking named entities in search engine queries. We steer the Semanticizer’s linking towards named entities by adapting an existing training corpus, and extend the Semanticizer’s set of features with contextual features that aim to leverage the limited context provided by search queries. With an F1 score of 0.6062 our final system run achieves median performance, and better than mean performance (0.5329).
References
-
Today marks my first day at Apple! I’ve joined the Apple Speech Recognition team within Siri, part of the AI/ML (Artificial Intelligence and Machine Learning) organization. After 12 wonderful years in Amsterdam, I’ve relocated to Cupertino, California, in the United States. I’m thrilled to be working at Apple Park, one of the most iconic and beautiful buildings in the world. Excited for the journey ahead!
-
Abstract
Modeling user browsing behavior is an active research area with tangible real-world applications, e.g., organizations can adapt their online presence to their visitors browsing behavior with positive effects in user engagement, and revenue. We concentrate on online news agents, and present a semi-supervised method for predicting news articles that a user will visit after reading an initial article. Our method tackles the problem using language intent models trained on historical data which can cope with unseen articles. We evaluate our method on a large set of articles and in several experimental settings. Our results demonstrate the utility of language intent models for predicting user browsing behavior within online news sites.