Matthias Sperber

Latest

• Automating Behavioral Testing in Machine Translation
• Towards Real-World Streaming Speech Translation for Code-Switched Speech
• Findings of the IWSLT 2023 Evaluation Campaign
• Joint Speech Transcription and Translation: Pseudo-Labeling with Out-of-Distribution Data
• Findings of the IWSLT 2022 Evaluation Campaign
• End-to-End Speech Translation for Code Switched Speech
• Parallelizing Autoregressive Neural Machine Translation via Modeling Half Sequences
• Simultaneous Speech Translation panel discussion at IWSLT
• Guest lecture at Maastricht University
• Streaming Models for Joint Speech Recognition and Translation
• Speech-translation tutorial at SOCML 2020
• Presented our TACL paper at EMNLP'20
• 2 speech translation posters at WeCNLP'20
• Streaming Models for Joint Speech Recognition and Translation
• Consistent Transcription and Translation of Speech
• Presented our theme track paper at ACL'20
• Speech Translation and the End-to-End Promise: Taking Stock of Where We Are
• Variational Neural Machine Translation with Normalizing Flows
• End-to-End Speech Translation - Opportunities & Challenges (colloquium talk @ Uni Heidelberg)
• End-to-End Speech Translation - Opportunities & Challenges (colloquium talk @ Uni Zurich)
• Some press coverage for our speech translation overview paper
• Spoken Language Translation (Interspeech Survey)
• Speech Translation (lecture @ MTM 2019)
• 4 papers at ACL'19 in Florence (3 of them on speech translation)
• Exploring Phoneme-Level Speech Representations for End-to-End Speech Translation
• Paraphrases as Foreign Languages in Multilingual Neural Machine Translation
• Self-Attentional Models for Lattice Inputs
• Fluent Translations from Disfluent Speech in End-to-End Speech Translation
• Attention-Passing Models for Robust and Data-Efficient End-to-End Speech Translation
• joined Apple as a machine translation researcher in Aachen
• defended my PhD on end-to-end speech translation
• End-to-End Neural Speech Translation
• KIT's IWSLT 2018 SLT Translation System
• Open Source Toolkit for Speech to Text Translation
• Low-Latency Neural Speech Translation
• Self-Attentional Acoustic Models
• XNMT: The eXtensible Neural Machine Translation Toolkit
• KIT's Multilingual Neural Machine Translation Systems for IWSLT 2017
• Toward Robust Neural Machine Translation for Noisy Input Sequences
• Transcribing Against Time
• Neural Lattice-to-Sequence Models for Uncertain Inputs
• Comparison of Decoding Strategies for CTC Acoustic Models
• Lightly Supervised Quality Estimation
• Adaptation and Combination of NMT Systems: The KIT Translation Systems for IWSLT 2016
• Audio Segmentation for Robust Real-Time Speech Recognition Based on Neural Networks
• The 2016 KIT IWSLT Speech-to-Text Systems for English and German
• Dynamic Transcription for Low-latency Speech Translation
• The Karlsruhe Institute of Technology Systems for the News Translation Task in WMT 2016
• Lecture Translator - Speech translation framework for simultaneous lecture translation
• Optimizing Computer-Assisted Transcription Quality with Iterative User Interfaces
• The 2015 KIT IWSLT Speech-to-Text Systems for English and German
• On-the-Fly User Modeling for Cost-Sensitive Correction of Speech Transcripts
• SESLA Transcriber: A Speech Transcription Tool That Adapts To Your Skill And Time Budget
• The 2014 KIT IWSLT Speech-to-Text Systems for English , German and Italian
• Segmentation for Efficient Supervised Language Annotation with an Explicit Cost-Utility Tradeoff
• The 2013 KIT IWSLT Speech-to-Text Systems for German and English
• Efficient Speech Transcription Through Respeaking
• The 2012 KIT and KIT-NAIST English ASR Systems for the IWSLT Evaluation
• The KIT-NAIST (Contrastive) English ASR Systems for the IWSLT Evaluation
• Efficient Transcription Through Respeaking
• JAGUC--a software package for environmental diversity analyses.
• Handwriting Reconstruction For a Camera Pen Using Random Dot Patterns
• A Camera Pen Using LLAH and Randomized Dot Patterns
• Position Detection For a Camera Pen Using LLAH and Dot Patterns