Text Stream
Text stream processing focuses on efficiently analyzing and extracting information from continuously arriving sequences of text data, a challenge posed by the ever-increasing volume of online textual information. Current research emphasizes developing models capable of handling the infinite length, sparsity, and evolving nature of these streams, often employing techniques like linear-time decoding, sliding window approaches, and incremental learning algorithms such as those based on SBERT and incremental SVMs. This field is crucial for real-time applications like zero-shot text-to-speech, LLM-based text streaming services, and sentiment analysis, improving user experience and enabling timely insights from dynamic textual data.
Papers
Methods for Generating Drift in Text Streams
Cristiano Mesquita Garcia, Alessandro Lameiras Koerich, Alceu de Souza Britto, Jean Paul Barddal
Improving Sampling Methods for Fine-tuning SentenceBERT in Text Streams
Cristiano Mesquita Garcia, Alessandro Lameiras Koerich, Alceu de Souza Britto Jr, Jean Paul Barddal
Frustratingly Easy Sentiment Analysis of Text Streams: Generating High-Quality Emotion Arcs Using Emotion Lexicons
Daniela Teodorescu, Saif M. Mohammad
JOIST: A Joint Speech and Text Streaming Model For ASR
Tara N. Sainath, Rohit Prabhavalkar, Ankur Bapna, Yu Zhang, Zhouyuan Huo, Zhehuai Chen, Bo Li, Weiran Wang, Trevor Strohman