Language models (LMs) often struggle to pay enough attention to the input context, and generate texts that are unfaithful or contain hallucinations. To mitigate this issue, we present context-aware decoding (CAD), which follows a contrastive output distribution that amplifies the difference between the output probabilities when a model is used with and without context. Our experiments show that CAD, without additional training, significantly improves the faithfulness of different LM families, including OPT, GPT, LLaMA and FLAN-T5 for summarization tasks (e.g., 14.3% gain for LLaMA in factuality metrics). Furthermore, CAD is particularly effective in overriding a model's prior knowledge when it contradicts the provided context, leading to substantial improvements in tasks where resolving the knowledge conflict is essential.

Current abstractive summarization systems present important weaknesses which prevent their deployment in real-world applications, such as the omission of relevant information and the generation of factual inconsistencies (also known as hallucinations). At the same time, automatic evaluation metrics such as CTC scores have been recently proposed that exhibit a higher correlation with human judgments than traditional lexical-overlap metrics such as ROUGE. In this work, we intend to close the loop by leveraging the recent advances in summarization metrics to create quality-aware abstractive summarizers. Namely, we propose an energy-based model that learns to re-rank summaries according to one or a combination of these metrics. We experiment using several metrics to train our energy-based re-ranker and show that it consistently improves the scores achieved by the predicted summaries. Nonetheless, human evaluation results show that the re-ranking approach should be used with care for highly abstractive summaries, as the available metrics are not yet sufficiently reliable for this purpose.

The series, which is produced by BBC StoryWorks Commercial Productions and presented by the International Science Council (ISC), includes films, articles and podcasts which will be hosted on a dedicated BBC.com StoryWorks webpage. The series explores how scientific culture is changing for the better, towards a future of more effective and inclusive citizen engagement, interdisciplinary and international cooperation, and open knowledge-sharing. This five-minute film highlights the innovative use of shipwrecks to map the seabed to inform the siting of offshore windfarms as the seas around Ireland provide an abundance of wind resources. Shipwrecks disturb near-seabed currents, causing certain types of sediments to be washed away or eroded. By studying these changes, we can better predict how man-made structures including wind turbines, will behave on the seabed over time.

It's karaoke-rehearsal time at Knollwood Military Retirement Community, a 300-bed facility tucked away in a leafy corner of northwest Washington, D.C. Knollwood resident and retired U.S. Army Colonel Phil Soriano, 86, has hosted the facility's semi-monthly singalongs since their debut during a boozy snowstorm happy hour in 2016. For the late August 2019 show, he'll share emcee duties with a special guest: Stevie, a petite and personable figure who's been living at Knollwood for the last six weeks. Soriano wants to sing the crowd-pleasing hit "YMCA" while Stevie leads the crowd through the song's signature dance moves. But Stevie is a robot, and this is harder than it sounds. "We could try to make him dance," says Niamh Donnelly, the robot's lead AI engineer, though she sounds dubious. She enters commands on a laptop.

It's karaoke-rehearsal time at Knollwood Military Retirement Community, a 300-bed facility tucked away in a leafy corner of northwest Washington, D.C. Knollwood resident and retired U.S. Army Colonel Phil Soriano, 86, has hosted the facility's semi-monthly singalongs since their debut during a boozy snowstorm happy hour in 2016. For the late August 2019 show, he'll share emcee duties with a special guest: Stevie, a petite and personable figure who's been living at Knollwood for the last six weeks. Soriano wants to sing the crowd-pleasing hit "YMCA" while Stevie leads the crowd through the song's signature dance moves. But Stevie is a robot, and this is harder than it sounds. "We could try to make him dance," says Niamh Donnelly, the robot's lead AI engineer, though she sounds dubious. She enters commands on a laptop.

Relational database management systems (RDBMSs) are powerful because they are able to optimize and answer queries against any relational database. A natural language interface (NLI) for a database, on the other hand, is tailored to support that specific database. In this work, we introduce a general purpose transfer-learnable NLI with the goal of learning one model that can be used as NLI for any relational database. We adopt the data management principle of separating data and its schema, but with the additional support for the idiosyncrasy and complexity of natural languages. Specifically, we introduce an automatic annotation mechanism that separates the schema and the data, where the schema also covers knowledge about natural language. Furthermore, we propose a customized sequence model that translates annotated natural language queries to SQL statements. We show in experiments that our approach outperforms previous NLI methods on the WikiSQL dataset and the model we learned can be applied to another benchmark dataset OVERNIGHT without retraining.