JHU HLTCOE was a top performer in a recent open speaker recognition challenge called VoxSRC, finishing in the top two of more than fifty entries from the international research community. The challenge, hosted by the University of Oxford, was based on the open-source VoxCeleb speech corpus captured from public celebrity videos with automatic speaker labeling […]
Videos include information in both the visual and audio domains, and so video processing techniques should utilize both of these means for more effective solutions. The HLTCOE has been researching this strategy since 2017, when recognizing individuals in videos using both voice and face was a topic at the SCALE summer workshop. The work resulting […]
Bilingual lexicons (or bilingual dictionaries) are valuable resources for machine translation. For example, when working with technical documents like patents, a bilingual lexicon consisting of technical jargon is important for ensuring that the translation is precise and correct. At the conference on Empirical Methods in Natural Language Processing (EMNLP) in November 2019, JHU researchers released […]
by Kevin Duh, Senior Research Scientist The SCALE2018 Machine Translation workshop focused on building resilient neural machine translation systems for new domains. In addition to developing new algorithms to improve translation accuracy, the team also dedicated significant efforts to analysis techniques in order to understand when and why neural networks work. Neural network models, […]
by Greg Sell, Senior Research Scientist A team of Johns Hopkins researchers from the HLTCOE and CLSP participated in the recent DIHARD challenge. In the evaluation, teams were given audio recordings of speech from a diverse set of conditions with an unknown number of speakers, with the goal of correctly marking the times that each […]
Cross-lingual information extraction is the task of distilling facts from foreign language (e.g. Chinese text) into representations in another language that is preferred by the user (e.g. English tuples). Conventional pipeline solutions decompose the task as machine translation followed by information extraction (or vice versa). We propose a joint solution with a neural sequence model, […]
by Kevin Duh, Senior Research Scientist The HLTCOE has acquired a NVIDIA DGX-1 Deep Learning System to support its numerous deep learning research efforts. With eight Tesla P100 GPUs, this brand new system delivers 170 teraflops of computing power (equivalent to 250 x86 servers) and has been described as an “AI Supercomputer in a […]
