Search for: All records

Transfer learning using ImageNet pre-trained models has been the de facto approach in a wide range of computer vision tasks. However, fine-tuning still requires task-specific training data. In this paper, we propose N3 (Neural Networks from Natural Language) - a new paradigm of synthesizing task-specific neural networks from language descriptions and a generic pre-trained model. N3 leverages language descriptions to generate parameter adaptations as well as a new task-specific classification layer for a pre-trained neural network, effectively “fine-tuning” the network for a new task using only language descriptions as input. To the best of our knowledge, N3 is the first method to synthesize entire neural networks from natural language. Experimental results show that N3 can out-perform previous natural-language based zero-shot learning methods across 4 different zero-shot image classification benchmarks. We also demonstrate a simple method to help identify keywords in language descriptions leveraged by N3 when synthesizing model parameters. 

Abstract Elton's biotic resistance hypothesis, which posits that diverse communities should be more resistant to biological invasions, has received considerable experimental support. However, it remains unclear whether such a negative diversity–invasibility relationship would persist under anthropogenic environmental change. By using the common ragweed (Ambrosia artemisiifolia) as a model invader, our 4‐year grassland experiment demonstrated consistently negative relationships between resident species diversity and community invasibility, irrespective of nitrogen addition, a result further supported by a meta‐analysis. Importantly, our experiment showed that plant diversity consistently resisted invasion simultaneously through increased resident biomass, increased trait dissimilarity among residents, and increased community‐weighted means of resource‐conservative traits that strongly resist invasion, pointing to the importance of both trait complementarity and sampling effects for invasion resistance even under resource enrichment. Our study provides unique evidence that considering species’ functional traits can help further our understanding of biotic resistance to biological invasions in a changing environment.