Current Challenges and Future Directions in Podcast Information Access

Jones, Rosie; Zamani, Hamed; Schedl, Markus; Chen, Ching-Wei; Reddy, Sravana; Clifton, Ann; Karlgren, Jussi; Hashemi, Helia; Pappu, Aasish; Nazari, Zahra; Yang, Longqi; Semerci, Oguz; Bouchard, Hugues; Carterette, Ben

doi:10.1145/3404835.3462805

Cited by 19 publications

(14 citation statements)

References 86 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…business, true crime). Podcasts have been massively popularised in the recent years due to increased use of streaming platforms and availability of underlying technology for information access, recording and publishing [11,20,35]. As of summer 2021, the number of available podcasts in the world exceeds 2M and over 48M episodes [32].…”

Section: Introductionmentioning

confidence: 99%

“…However, annotating podcasts with topics is still quite problematic. First, although podcast metadata entails topic-related genres, manually assigned by creators or providers, in reality these are often noisy and unreliable [11,28]: genres could be too heterogeneous or broad (e.g. Kids & Family includes both sleep meditation and parenting advice); and podcast providers may misleadingly label their shows with unrelated genres for broadening exposure.…”

Section: Introductionmentioning

confidence: 99%

“…Kids & Family includes both sleep meditation and parenting advice); and podcast providers may misleadingly label their shows with unrelated genres for broadening exposure. Second, using topic modeling has its limitations inherited from the input text associated with podcasts: 1) metadata, such as title or description, is typically a short text of varying quality; 2) the automatically transcribed speech is noisy having a high word-error-rate especially for NEs and requires the engagement of more resources [3,8,11].…”

Section: Introductionmentioning

confidence: 99%

“…Identifying podcast categories at different granularity levels could help editors evolve manually created podcast taxonomies and automatically annotate podcasts with these categories. The discovered topics could also support the consolidation of podcast knowledge graphs [2,11], recently exploited in recommendation, by adding new edges that capture topic-informed podcast similarity based on metadata.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Topic Modeling on Podcast Short-Text Metadata

Valero¹,

Baranes²,

Epure³

2022

Preprint

View full text Add to dashboard Cite

Podcasts have emerged as a massively consumed online content, notably due to wider accessibility of production means and scaled distribution through large streaming platforms. Categorization systems and information access technologies typically use topics as the primary way to organize or navigate podcast collections. However, annotating podcasts with topics is still quite problematic because the assigned editorial genres are broad, heterogeneous or misleading, or because of data challenges (e.g. short metadata text, noisy transcripts). Here, we assess the feasibility to discover relevant topics from podcast metadata, titles and descriptions, using topic modeling techniques for short text. We also propose a new strategy to leverage named entities (NEs), often present in podcast metadata, in a Non-negative Matrix Factorization (NMF) topic modeling framework. Our experiments on two existing datasets from Spotify and iTunes and Deezer, a new dataset from an online service providing a catalog of podcasts, show that our proposed document representation, NEiCE, leads to improved topic coherence over the baselines. We release the code for experimental reproducibility of the results 1 .

show abstract

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Topic Modeling on Podcast Short-Text Metadata

Valero¹,

Baranes²,

Epure³

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Podcasts are very popular with internet users today rather than listening to broadcasts on the radio that have limitations in access, frequency, and time (Chan-Olmsted & Wang, 2022). Podcasts have been widely used primarily in the entertainment and education industries (Jones et al, 2021). The Guardian journalist Ben Hammersley 2004 proposed the term "podcast" which stands for "play-on-demand" and "broadcast".…”

Section: Introductionmentioning

confidence: 99%

The Role of Youth in the Stunting Prevention via Podcast in Karawang Regency 2020

Rahayu

Apriningrum

et al. 2022

ISS

View full text Add to dashboard Cite

Podcasts are audio recordings that can be listened to by the general public through internet media. Unlike the radio which must be done and broadcast live at a certain frequency. Podcasts can be used as a medium of information related to stunting podcast management by teenagers is expected to better reach stunting prevention goals appropriately. This study aims to provide knowledge about stunting prevention and information dissemination skills about stunting prevention through podcasts in adolescents. The question and answer lecture method is used in the presentation of knowledge, while podcast training uses the demonstration method. This study was conducted from August to October 2020, with a number of respondents 10 adolescent participants. Through this service activity, increasing adolescent knowledge about stunting and spreading stunting information through podcasts to the knowledge of all respondents in the good category (100%), adolescent experience and skills in podcast management became quite 10% and good 90%.

show abstract