Summarizing multiple disaster-relevant data streams simultaneously is particularly challenging as existing Retrieve&Re-ranking strategies suffer from the inherent redundancy of multi-stream data and limited scalability in a multi-query setting. This work proposes an online approach to crisis timeline generation based on weak annotation with Deep Q-Networks. It selects on-the-fly the relevant pieces of text without requiring neither human annotations nor content re-ranking. This makes the inference time independent of the number of input queries. The proposed approach also incorporates a redundancy filter into the reward function to effectively handle cross-stream content overlaps. The achieved ROUGE and BERTScore results are superior to those of best-performing models on the CrisisFACTS 2022 benchmark.

DQNC2S: DQN-based Cross-stream Crisis event Summarizer / Rege Cambrin, Daniele; Cagliero, Luca; Garza, Paolo. - 14610:(2024), pp. 422-430. (Intervento presentato al convegno 46th European Conference on Information Retrieval. ECIR 2024 tenutosi a Glasgow (UK) nel March 24–28, 2024) [10.1007/978-3-031-56063-7_34].

DQNC2S: DQN-based Cross-stream Crisis event Summarizer

Rege Cambrin, Daniele;Cagliero, Luca;Garza, Paolo
2024

Abstract

Summarizing multiple disaster-relevant data streams simultaneously is particularly challenging as existing Retrieve&Re-ranking strategies suffer from the inherent redundancy of multi-stream data and limited scalability in a multi-query setting. This work proposes an online approach to crisis timeline generation based on weak annotation with Deep Q-Networks. It selects on-the-fly the relevant pieces of text without requiring neither human annotations nor content re-ranking. This makes the inference time independent of the number of input queries. The proposed approach also incorporates a redundancy filter into the reward function to effectively handle cross-stream content overlaps. The achieved ROUGE and BERTScore results are superior to those of best-performing models on the CrisisFACTS 2022 benchmark.
2024
978-3-031-56062-0
978-3-031-56063-7
File in questo prodotto:
File Dimensione Formato  
ECIR2024.pdf

embargo fino al 23/03/2025

Tipologia: 2. Post-print / Author's Accepted Manuscript
Licenza: PUBBLICO - Tutti i diritti riservati
Dimensione 419.18 kB
Formato Adobe PDF
419.18 kB Adobe PDF   Visualizza/Apri   Richiedi una copia
978-3-031-56063-7_34.pdf

non disponibili

Tipologia: 2a Post-print versione editoriale / Version of Record
Licenza: Non Pubblico - Accesso privato/ristretto
Dimensione 276.36 kB
Formato Adobe PDF
276.36 kB Adobe PDF   Visualizza/Apri   Richiedi una copia
Pubblicazioni consigliate

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11583/2984812