A Robustly Optimized BERT Pre-training Approach with Post-training

We improved the quality and reduced the time to produce machine learned models for use in small molecule antiviral design. Our globally asynchronous multi-level parallel training approach strong scales to all of Sierra with up to 97.7% efficiency. We trained a novel, character-based Wasserstein autoencoder that produces a higher quality model trained on 1.613 billion compounds in 23 minutes while the previous state of the art takes a day on 1 million compounds. Reducing training time from a day to minutes shifts the model creation bottleneck from computer job turnaround time to human innovation time. Our implementation achieves 318 PFLOPs for 17.1% of half-precision peak. We will incorporate this model into our molecular design loop enabling the generation of more diverse compounds; searching for novel, candidate antiviral drugs improves and reduces the time to synthesize compounds to be tested in the lab.

Download Full-text

Beyond Planning Tools: Experiential Learning in Climate Adaptation Planning and Practices

Climate ◽

10.3390/cli9050076 ◽

2021 ◽

Vol 9 (5) ◽

pp. 76

Author(s):

Kristen M. Schmitt ◽

Todd A. Ontl ◽

Stephen D. Handler ◽

Maria K. Janowiak ◽

Leslie A. Brandt ◽

...

Keyword(s):

Natural Resources ◽

Climate Adaptation ◽

Climate Science ◽

Adaptation Planning ◽

Work Time ◽

Training Approach ◽

Climate Change Response ◽

Planning Tools ◽

Change Response

In the past decade, several dedicated tools have been developed to help natural resources professionals integrate climate science into their planning and implementation; however, it is unclear how often these tools lead to on-the-ground climate adaptation. Here, we describe a training approach that we developed to help managers effectively plan to execute intentional, climate-informed actions. This training approach was developed through the Climate Change Response Framework (CCRF) and uses active and focused work time and peer-to-peer interaction to overcome observed barriers to using adaptation planning tools. We evaluate the effectiveness of this approach by examining participant evaluations and outlining the progress of natural resources projects that have participated in our trainings. We outline a case study that describes how this training approach can lead to place and context-based climate-informed action. Finally, we describe best practices based on our experience for engaging natural resources professionals and helping them increase their comfort with climate-informed planning.

Download Full-text

A Self-Training Approach for Point-Supervised Object Detection and Counting in Crowds

IEEE Transactions on Image Processing ◽

10.1109/tip.2021.3055632 ◽

2021 ◽

Vol 30 ◽

pp. 2876-2887

Author(s):

Yi Wang ◽

Junhui Hou ◽

Xinyu Hou ◽

Lap-Pui Chau

Keyword(s):

Object Detection ◽

Training Approach ◽

Detection And Counting

Download Full-text

A Structured-Training Approach to Teaching Diagnostic Interviewing

Teaching of Psychology ◽

10.1177/009862839402100304 ◽

1994 ◽

Vol 21 (3) ◽

pp. 144-150 ◽

Cited By ~ 1

Author(s):

Susan M. Bögeis

Keyword(s):

Mental Health ◽

Health Professions ◽

Simulated Patients ◽

Training Methods ◽

The Past ◽

Training Approach ◽

Interviewing Skills ◽

Recent Developments ◽

Approach To Teaching ◽

Clinical Problems

In this article, recent developments in the teaching of diagnostic-interviewing skills in the mental health professions are discussed. First, the tasks and skills of the diagnostic interviewer are analyzed. Second, empirical evidence on effective interviewing styles is outlined. Third, training methods for teaching diagnostic interviewing are reviewed. A training program, developed to teach diagnostic interviewing to undergraduates, is then described. The program is highly structured, and simulated patients are used to introduce complex clinical problems and to evaluate students’ emerging competencies. Student and trainer satisfaction with the program has been high during the past 6 years.

Download Full-text