Delhi Development Authority

Both long-tailed and noisily labeled data frequently appear in real-world applications and impose significant challenges for learning. Most prior works treat either problem in an isolated way and do not explicitly consider the coupling effects of the two. Our empirical observation reveals that such solutions fail to consistently improve the learning when the dataset is long-tailed with label noise. Moreover, with the presence of label noise, existing methods do not observe universal improvements across different sub-populations; in other words, some sub-populations enjoyed the benefits of improved accuracy at the cost of hurting others. Based on these observations, we introduce the Fairness Regularizer (FR), inspired by regularizing the performance gap between any two sub-populations. We show that the introduced fairness regularizer improves the performances of sub-populations on the tail and the overall learning performance. Extensive experiments demonstrate the effectiveness of the proposed solution when complemented with certain existing popular robust or class-balanced methods.

Intensifying tropical cyclones threaten linked socio-ecological systems. The Philippine abaca (Musa textilis), comprising >85% of global supply, is concentrated in the world's most typhoon-prone landscapes, yet its coupled vulnerability with endemic forest biodiversity remains unquantified. We develop a novel framework to map where climate hazard, livelihood dependence, and conservation priority intersect. Using 45 years of typhoon data (1980-2025), we integrate metrics for abaca cultivation and forest biodiversity to create the first high-resolution Co-Vulnerability Index for any Philippine island. We reveal that the highest-risk zones are concentrated within protected area networks. Here, typhoon exposure drives 48% of the risk, followed by biodiversity value (33%) and abaca livelihoods (20%). This critical finding, that formal protection coincides with peak vulnerability, challenges the view of protected areas as passive refuges. It instead underscores their necessary role as active platforms for integrated climate adaptation, providing a scalable targeting tool for cyclone-prone regions worldwide.

Abstract. In India, land ownership disputes remain a major source of litigation, often resulting from fragmented records, lack of transparency, and outdated administrative processes. This study explores the potential of geospatially enabled blockchain technology to transform land administration into a secure, efficient, and trust based system. By combining the immutability and decentralization of blockchain with the spatial precision of geospatial data, the research proposes a digital framework that enhances the integrity of land transactions and builds stakeholder confidence. The methodology follows a modular system development approach using both software and database life cycle models. Through fact finding, stakeholder interviews, prototyping, and interface development, the system was designed to accommodate various users buyers, sellers, inspectors, and administrators. Key features include cryptographic public private key generation and secure hashing algorithms for transaction protection. A functional prototype was developed using Java and MySQL, integrating GIS capabilities and smart contracts to automate processes and ensure spatial traceability. The system’s performance was assessed based on its completeness, usability, and security. Results showed that the blockchain ledger enabled real time synchronization and minimized the risk of fraud, tampering, or duplication. Incorporating spatial data allowed for visual verification of land parcels and ownership histories, improving reliability for both citizens and government agencies. This research demonstrates how integrating geospatial intelligence with blockchain can modernize land governance, reduce bureaucratic delays, and foster a more transparent and citizen centric system. It presents a promising direction for high performance, scalable digital infrastructure in rapidly developing regions.

Key Clinical Message: Lupus cerebritis presents diagnostic and therapeutic challenges.Refractory seizures may require early aggressive immunosuppression.Emerging biomarkers like TWEAK offer promise in guiding management.Multidisciplinary care and vigilant follow-up are essential to reduce neuropsychiatric morbidity and improve outcomes in neuropsychiatric lupus.

The spatial distribution of food shapes animal movement patterns, enabling them to optimize search behaviour and establish efficient routes for resource exploitation. Anthropogenic alteration of primate habitats, fragments available resources, and increases dependency on human food. The rhesus macaque (Macaca mulatta) exhibits a broad spectrum of habitat preferences, encompassing both natural and human-altered landscapes. We studied how these preferences influence ranging behaviour in synanthropic macaques. We studied two similar-sized synanthropic rhesus macaque groups in a 16 km2 suburban area in Dehradun, India. We tagged an adult female from each group with a GPS-GSM logger, recording fixes at 15-min intervals during the summer (April–June) and monsoon (July–August) of 2022. We investigated ranging patterns, habitat selection (via integrated step selection analysis, which models animal movement and habitat preference by analysing step choices based on current location and habitat characteristics), site fidelity, and movement strategy (via evaluation of consecutive step lengths) of the macaques during these two seasons based on the logged GPS coordinates during this period. Overall, the home and core ranges were small. The mean total distance covered diurnally was ~5–8 times the mean of the shortest distance between the first and last location diurnally. The macaques showed a strong preference for roadside and human habitation. Both individuals showed high site fidelity. Both individuals used a strategy of short intensive searches followed by long directed movements. The preference for anthropogenic sites and site fidelity along habitat edges shows recurrent movement to human dumpsites and houses, followed by return trips to preferred refuges. The overall movement strategy of these macaques in human-modified landscapes involved travel between clustered foraging and resting sites, reflecting an adaptive response to spatially fragmented resources and highlighting how resource distribution influences search efficiency and route optimization. This movement pattern may reinforce perceptions of heightened agonistic interactions among humans and influence the long-term dynamics of human-macaque coexistence.