7627 research outputs found
Sort by
Towards scalable taxi demand prediction
Accurate taxi demand prediction is essential for optimizing urban mobility systems across varying spatial-temporal resolutions and data conditions. Scalable taxi demand prediction refers to the capability of forecasting models to adapt to different granularities of spatial and temporal data while maintaining prediction accuracy, a critical requirement for practical urban applications ranging from fleet management to transportation planning. However, two fundamental challenges impede this scalability: data sparsity and multi-resolution forecasting requirements. Data sparsity, particularly pronounced in high-resolution predictions where numerous regions exhibit minimal activity, significantly compromises model performance. Concurrently, different urban applications necessitate predictions at varying temporal and spatial granularities, requiring models with adaptive forecasting capabilities. Our research systematically addresses these challenges by first developing solutions for data sparsity, which then enable effective multi-resolution prediction capabilities. We introduce STZIP-GNN, a Spatiotemporal Zero- Inflated Poisson Graph Neural Network that explicitly models structural zeros through a ZIP probability layer while integrating auxiliary information from crowdsensed geolocation and socioeconomic data. This approach effectively mitigates the sparse demand patterns frequently encountered in high-resolution data. Building upon this foundation, we develop SSTZIP-GNN, which incorporates an adaptive learning mechanism that dynamically adjusts forecasting strategies based on input data granularity, enabling scalable, multi-resolution prediction within a unified framework. Extensive experiments conducted on 130 million real-world mobility records demonstrate that our proposed models significantly outperform existing baselines. SSTZIP- GNN achieves up to 53.3% RMSE reduction and 46.3% lower computational costs, exhibiting superior robustness and scalability across diverse urban conditions. These advancements provide municipalities with a unified tool for demand-responsive fleet management, dynamic pricing, and sustainable mobility planning across heterogeneous urban landscapes. Our research contributes to the development of more efficient and adaptable intelligent transportation systems
Three-stage latent dynamics forecasting (T-LDF) framework for Shenzhen Metro passenger flow prediction
Accurate forecasting of metro passenger flow is vital for efficient urban transportation management and optimal resource allocation in modern cities. Traditional ARIMA-based models effectively capture regular, cyclical patterns but struggle with sudden, nonlinear fluctuations caused by random events such as weather disruptions, special events, or service interruptions. Moreover, existing research predominantly focuses on individual stations, overlooking the complex cross-station interactions inherent in networked metro systems where passenger flows are interconnected across the entire network.
To address these critical limitations, we propose the Three-Stage Latent Dynamics Forecasting (T-LDF) Framework, a novel approach that systematically integrates temporal decomposition, latent dynamics extraction, and cross-station interaction modeling. Initially, the observed passenger flow is decomposed into a periodic component, representing regular daily and weekly commuting patterns, and a non-periodic component, capturing irregular fluctuations and anomalies. Next, the Latent Variable AutoRegression (LaVAR) model extracts low-dimensional dynamic latent variables (DLVs) from the non-periodic component, effectively filtering noise while preserving essential dynamic information for model training. Finally, a supervised learning model with an innovative alignment mechanism accounts for travel delay effects in passenger movements across the metro network, facilitating accurate real-time forecasting. Comprehensive experiments conducted on the Shenzhen Metro passenger flow dataset demonstrate that T-LDF significantly outperforms traditional statistical models and existing single-station approaches, achieving remarkable Test-RMSE reductions of 27% on weekdays and 20.4% on weekends, while simultaneously enhancing model interpretability through explicit cross-station interaction modeling
從香港填海歷史看地名之名實變遷
地名與生活處處相關,然香港自19世紀以來的大規模填海工程,雖然幫助本港增加大量用地,但亦因亦改變了城市的地理面貌,也深刻影響了地名的「名實」。本研究以填海歷史為主線,循古到今,系統分析海灣、岬角、島嶼及水系四類地貌的地名變遷,梳理地貌改變時間,探討填海如何導致地名與實際地形脫節,並揭示地名背後的社會文化意義。研究發現,填海工程對地名的影響主要呈現三種模式:自然地貌消失但地名沿用、地名範圍隨填海擴大、地名隨填海消失。研究進一步發現地名變遷存在權力介入及民間力量。
本研究填補了香港地名研究中填海影響的空白,更結合地圖、文獻與民間資料,呈現地名作為歷史載體的意義。儘管填海工程使地貌「面目全非」,地名仍成為市民追溯昔日文化的重要線索
“塵外塤箎”與中日詩誼 : 《元元唱和集》新探
江户前期,東渡日本的浙江籍文人陳元贇與日蓮宗僧元政詩歌贈答往來,唱和盈帙並合編爲《元元唱和集》。該集是東亞漢文學史上的第一部中日人士唱和專集,既見證了士僧之間神交心契、塤篪相應的深厚友誼,又呈示了明清嬗替時期與江户漢文學交流互動的多元樣態和詩藝圖景。書翰往來與詩歌唱和交際功能的耦合,不僅在文學場域產生了雙重疊加效應,同時也俾使“元元唱和”活動得以跨時空延續。“元元”唱和詩歌在抒寫日常生活、自然四時、行旅思歸與讚頌親友之情等層面,多向度地映現了中日漢詩的同源異流特徵。作爲近世中日文化交流史上的詩誼典範,通過對“元元唱和”及唱和集内容的尋繹,可覘其中含藴著豐富的文學文化價值。
In the early Edo period, Chen Yuanyun (陳元贇, 1587-1671), a scholar from Zhejiang who traveled to Japan, exchanged poems with Gensei (元政, 1623-1668), a monk of the Nichiren sect (日蓮宗), and their poems were compiled into Yuan Yuan changhe ji (《元元唱和集》). This collection is the first special collection of Chinese and Japanese poetry and poetry in the history of East Asian Chinese literature. It not only bears witness to the deep friendship between a scholar and a monk, but also presents the diverse forms and poetic landscape of exchanges and interactions with Edo Chinese literature during the transition period from Ming to Qing dynasties. The coupling of correspondence and the communicative functions of poetry not only produced a double superposition effect in the literary field, but also enabled the “Yuan Yuan changhe” (元元唱和) activity to continue across time and space. The “Yuan Yuan changhe” poems reflect the common origins and different streams of Chinese and Japanese poetry in multiple dimensions, such as expressing daily life, the four seasons in nature, homesickness of travelers, and praising the feelings of relatives and friends. As a model of poetic friendship in the history of cultural exchanges between China and Japan in modern times, through the exploration of the contents of “Yuan Yuan changhe” and the collection of changhe, we can see that it contains rich literary and cultural value
多元性視野下謝靈運的精神世界
中古貴族文士往往具有多元的知識與思想背景,謝靈運博學多才,出自家學濶深的陳郡謝氏家族,又通達儒學、佛教、道教及老莊玄學等,本爲時代思潮之前沿人物。但謝靈運本人的精神世界始終複雜多端,頗由外界因緣驅動而起伏興變。這一特點投射於作品文本中,亦呈現出多元交織卻無以歸宗的思想風貌。
Aristocratic literati of the medieval period often possessed diverse knowledge and intellectual backgrounds. Xie Lingyun, a highly learned and talented individual, hailed from the prestigious Xie clan of Chen Prefecture. He was well-versed in Confucianism, Buddhism, Daoism, and the metaphysical teachings of Laozi and Zhuangzi, making him a figure at the forefront of the intellectual trends of his time. However, Xie Lingyun’s spiritual world remained complex and multifaceted, often shaped and influenced by external circumstances, leading to fluctuations and transformations. This characteristic is reflected in his literary works, which exhibit a diverse and interwoven yet unanchored intellectual style
北周尉遲迥相州舉兵的多重敘事構造 : 以《周書.尉遲迥傳》《尉遲迥廟碑》及《集神州三寶感通録》爲中心
周隋革命的前夜,楊堅在帝都長安較爲順利地清除了異己,但在北周宣帝去世之後的三個月内,地方上先後有三支北周勤王之兵,即:相州總管尉遲迥、益州總管王謙、鄖州總管司馬消難。其中對於楊堅來説,最大的威脅來自鎮守東方的尉遲迥。本文詳細考察三類文獻:中古正史(《周書.尉遲迥傳》)、碑刻文獻(《尉遲迥廟碑》)、佛教史(《集神州三寳感通録》)對北周大象二年(580)六月尉遲迥相州舉兵這一事件的不同敘事。從北周到唐代,這一事件具有多重的敘事構造,每一種敘事構造的背後,體現了不同書寫者或者敘述者的需求與認同。
最值得留意的是唐代高僧道宣《集神州三寳感通録》,他通過對尉遲迥舉兵失敗之後的佛寺“感通”敘事,折射出了佛教史學與中古正统國史(碑刻)之間的内在張力,這種張力導致了多重敘事構造間的巨大差異。
On the eve of revolution between the Northern Zhou and the Sui Dynasty, Yang Jian eliminated the dissidents in the imperial capital Chang’an smoothly, but within three months after the death of Emperor Xuan of the Northern Zhou Dynasty, three armies which aimed at protecting the emperor were organized at the local level: Yuchi Jiong, the governor of Xiangzhou; Wang Qian, the governor of Yizhou; and Sima Xiaonan, the governor of Yunzhou. Among them, the biggest threat to Yang Jian came from Yuchi, who was guarding the east. This paper examines different narratives of Yuchi Jiong’s rebellion in Xiangzhou in June of the second year of the Daxiang period (580) in the Northern Zhou Dynasty, which are contained in three different documents: medieval official history (Biography of Yuchi Jiong in the History of Zhou), stone inscription (Yuchi Jiong’s Temple Tablet), and Buddhist history (Records of the Numinous Communications about the Three Jewels in China). From the Northern Zhou to the Tang Dynasty, the narrative constructions of this event are multiple, each of which implies the need and identification of different writers or narrators.
The most noteworthy document is Daoxuan’s Records of the Numinous Communications about the Three Jewels in China, which contains numinous narratives of Buddhist temples after Yuchi Jiong’s failure. These narratives reflect the inherent tension between Buddhist history and medieval official history (inscriptions), which lead to great discrepancies between multiple narrative constructions
“據譯單重”到多軌並途 : 漢文佛教目録分類之演變
典籍的分類方式,是佛教目録最重要的表現形式,也是佛教目録撰作者學術思想、觀念與撰録因由最爲直接的體現。漢文佛教目録最初本爲記録經典傳譯,於是產生了根據經典產生方式的“據譯單重”分類法。其後随著認識的深入,根據經典性質分爲經、律、論三藏,根據經典體制分爲大、小乘的“乘藏分類法”逐漸取得统治地位。明代末年,出現了依據天台五時判教分類經典的“依義銓次”分類法。近現代以來,由於西方學術觀念的傳入與知識生產方式的變化,漢文佛教目録的分類方式日趨多元。
The method of categorizing scriptures is the most crucial feature of Buddhist catalogs, directly reflecting the academic thought, perspectives, and motivations of their compilers. Initially, Chinese Buddhist catalogs were created to document the translation of scriptures, leading to the emergence of the “single-layer translation-based classification” method, which categorized texts based on their origins in translation. As understanding deepened, the “Tripiṭaka classification method” became dominant, dividing scriptures into the three categories of sutras (teachings), vinayas (disciplinary rules), and śāstra (treatises) based on their nature, and further into Mahāyāna and Hīnayāna texts according to their doctrinal frameworks. By the late Ming dynasty, a new “classification by doctrinal hierarchy” emerged, based on the Tiantai school’s Five Periods of Teachings system. In the modern era, with the influence of Western academic concepts and changes in knowledge production methods, the classification methods of Chinese Buddhist catalogs have become increasingly diverse
論東周銅器銘文仿古的文體學意義
東周器銘的仿古是立足新的時代語境而模仿古代文辭的一種書寫方式,是在純文本寫作日益佔據主導優勢的情況下,依然仿效“記録一修改”模式的某些特徵的做法。具體而言,仿古主要包括擬辭和擬篇兩個層面。前者是從古老的“詩”“書”類文獻和舊器銘文中採擇詞彙、語法對新文本進行“作舊”處理。後者是在西周器銘常見類型,如册命、訓誥等基本消隱的情況下,重又參照舊文本改寫事實以製作同類銘文。這造就了東周銘體獨特的演變節奏。仿古的主要目的在於利用銘體穩定的文本特性和禮制屬性以彰顯作器者的權力合法性。其影響經由傳世東周文獻的傳導,開啓了秦漢以後長期追求簡古弘深的銘體創作,對銘體傳统的塑造作用是關鍵的。其手法與原因,一定程度上又含有超出東周時代的啓發價值,具有特殊的文體學意義。
The Eastern Zhou Dynasty’s imitation of ancient bronze inscriptions was a way of writing that imitated the ancient languages based on a new era context. It was a practice that disregarded the increasing dominance of pure text writing and imitated features of the “record-modify” mode. It mainly included two levels of practical techniques: imitation of words and imitation of texts. The former involved selecting vocabulary and grammar from documents related to Shi and Shu, as well as inscriptions on old bronzes, to “nake old” the new texts. The latter was the practice of rewriting facts based on old texts to create similar inscriptions, disregarded the almost disappearance of common types of Western Zhou inscriptions such as Ming, Xun and Gao. This led to a unique evolution rhythm of the inscriptions during the Eastern Zhou Dynasty. The main purpose of the imitation was to utilize the stable textual characteristics and ritual attributes of inscriptions to demonstrate the legitimacy of the makers’ power. And its influence was transmitted through the literature of the Eastern Zhou Dynasty, which began the long-term pursuit of simple and elegant inscriptions after the Qin and Han dynasties. The Eastern Zhou Dynasty’s imitation of ancient bronze inscriptions played a crucial role in shaping the tradition of inscriptions, and its techniques and reasons, to a certain extent, also contained inspirational value that transcended the era, which demonstrated unique significance in genre studies
Promoting an expanded notion of decent work to enhance the career capabilities of young people
Young people, particularly those from impoverished backgrounds, face significant contextual challenges in their career and life development. This study, grounded in the capability approach and recognition theory, explores the expansion of decent work concepts to foster enabling working environments.
This seminar will address three key questions, supported by empirical evidence:1. What are the contextual challenges young people face in their career and life development? 2. What is the concept of decent work, and why is there a need to expand its scope? 3. How can recognition theory and the capability approach be integrated to inform the expansion of the decent work notion?
Finally, the seminar will discuss how a redefined notion of decent work, incorporating a recognition component, can support youth development in organizational contexts
दाइ मिङ: समय र विरासतको रखवाला
यो चित्र पुस्तक छापा र संरक्षण: प्रविधिभन्दा बाहिरका चिनियाँ पुस्तकहरू: अन्तरविषय अनुसन्धान परियोजनालाई पूरक बनाउन शैक्षिक पहुँच कार्यक्रमको एक भागको रूपमा प्रकाशित गरिएको छ। मनोहर कथाहरू मार्फत, बालबालिकाहरूले काठको ब्लक प्रिन्टिङ(मुद्रण) देखि डिजिटल मुद्रणसम्म, चिनियाँ पुस्तक मुद्रण प्रविधिको विकास र विविधताको अन्वेषण गर्न सक्छन् र वर्तमानदेखि विगतसम्म बुझ्न सक्छन्।आशा छ यो चित्र पुस्तक बालबालिकाहरूका लागि चिनियाँ पुस्तकहरू र मुद्रणको अद्भुत यात्रा सुरु गर्न सुरुवात बिन्दु बनोस्!https://commons.ln.edu.hk/picturebook_daiming/1002/thumbnail.jp