Search

Urban parks are one of the most common spaces for social interactions in modern cities. The design of park spaces, especially space configuration, has significant influences on people’s social behaviors in parks. In this study, the associations between space configurational attributes and social interactions were investigated using space syntax theory. An observation analysis of social behaviors was carried out in two urban parks in Beijing, China. Nine space configurational attributes, including depth to the gate, depth to the main road, connectivity, normalized angular integration (NAIN), and normalized angular choice (NACH) with three radii, were calculated using a segment model. The variance analysis and regression analysis reveal the strong joint effect of space type, space scale factors, and space configurational attributes on social interaction behaviors in parks. The personal interaction group contained 23% of the total observed people involved in social interactions. Pathway length, zone area, and NACH-10K (NACH with a radius of 10,000 m) are positively associated with the number of people involved in personal interactions. For the social interaction group (77% of the total observed people), the space scale and depth to main city road were found to have a positive and negative influence on social interaction intensity.

We present a neural model for paraphrasing and train it to generate delexicalized sentences. We achieve this by creating training data in which each input is paired with a number of reference paraphrases. These sets of reference paraphrases represent a weak type of semantic equivalence based on annotated slots and intents. To understand semantics from different types of slots, other than anonymizing slots, we apply convolutional neural networks (CNN) prior to pooling on slot values and use pointers to locate slots in the output. We show empirically that the generated paraphrases are of high quality, leading to an additional 1.29% exact match on live utterances. We also show that natural language understanding (NLU) tasks, such as intent classification and named entity recognition, can benefit from data augmentation using automatically generated paraphrases.

Street canyons are basic components of modern cities. Sound propagation in streets has been proven to be significantly affected by reflections from building fa&ccedil, ades. In this paper, an iterative model combining acoustic radiosity and the image source method (IMCRI) is proposed to investigate the effect of diffuse reflections on sound propagation in urban street canyons. By applying image patches, this model calculates both specular reflected components and diffuse reflected components in every reflection. The accuracy of this model is then validated with full-scale measured data from five actual streets with different scattering conditions. Good agreements of sound pressure level and reverberation time are found between the IMCRI and the measurements. The performance of the IMCRI is found to be superior to those of the existing energy models, especially for the reverberation time simulation.

This paper aims to find a series of objective indicators that can fully describe the soundscapes\ud of urban shopping streets. The perceptual and physical features of soundscapes in urban\ud shopping streets have been investigated via field surveys and a laboratory study. Using the\ud semantic differential method, the perception structure of shopping street soundscapes was\ud initially analysed, and five major perceptual factors were identified that explained 64% of the\ud total sample variance, including preference, loudness, communication, playfulness and\ud richness. Each perceptual factor explained approximately 10% to 15% of the total variance,\ud which showed that they shared very similar importance in the representation of the overall\ud soundscapes in urban shopping streets. Based on this semantic differential analysis, a\ud laboratory study was performed to investigate the relationship between perceptual factors and\ud physical indicators. Sound levels and psychoacoustic indicators were initially considered to be\ud the most common indicators. The results showed that there were significant correlations\ud between these indicators and three perceptual factors, including preference, loudness and\ud communication. A new indicator (dynamic spectrum centre or DSC) was then developed in\ud this study based on the concepts of spectrum gravity centre analysis and temporal variety\ud analysis. This indicator combined spectral and temporal analysis to describe the contents of\ud background sounds and sound events. The DSC indicator was found to have a significant\ud correlation with the other two perceptual factors (playfulness and richness). Increasing the\ud variability of the sound event (a higher standard deviation of DSC_E) could make soundscape\ud more playful and richer.