Indian Institute of Technology Gandhinagar

For someone who teaches environmental humanities and studies, and the human dimensions of wildlife conservation, living with ecological guilt is a daily reality. Constantly assessing one’s lifestyle and its environmental impact has been unavoidable. And the more materially comfortable life becomes, the sharper this guilt feels. I debate my use of plastic bags or cars for short distances almost daily. Cycling on campus offers momentary relief, but feels superficial. Riding a bicycle on a green, low-traffic campus is itself a privilege— one that quickly exposes the limits of individual virtue. In India, the ability to rethink lifestyle choices is, in truth, a luxury. What, then, of those who have no choice but to endure climate change, which is already eroding their health, livelihoods and, at times, their lives? Greater awareness of how both humans and non-humans are coping with climate disruptions only deepens the anxiety. Knowledge does not soothe; it compounds concern.

In this paper, we present an efficient deep learning architecture for real-world smartphone image denoising. The architecture is developed based on attention mechanism in a multi-patch hierarchical network with non-local Information. Unlike traditional methods that struggle with the spatially variant and complex noise patterns in smartphone images, our model integrates a multi-patch hierarchical approach to effectively leverage spatial context at multiple scales. The proposed network incorporates a non-local module in the encoder to capture long-range dependencies, enhancing the network’s ability to model global image structure. To further refine feature representation, we introduce a parallel attention mechanism in the decoder that combines both channel attention (CA) and pixel attention (PA). This dual attention design enables the network to emphasize relevant features both across channels and within local spatial regions, leading to improved denoising performance. Trained on real-world datasets, including spatially variant noise from smartphones, our method demonstrates superior quantitative and qualitative results while maintaining a lightweight architecture with fewer parameters. Experimental evaluations validate the effectiveness of our model.

Adaptive cueing in currently-existing technology-enhanced learning (TEL) often prioritizes performance metrics over the cognitive load experienced by a learner, thereby losing comprehensive view to realizing learning outcomes. Cues e.g., visual, auditory, tactile cues have been shown to be effective in guiding the learner's attention and managing cognitive load. However, offering cues to promote one's skill learning based only on individual performance indicators can be misleading, since added to the performance scores, cognitive aspects and learning ability of the learner are also crucial for effective learning. Though the cognitive aspects are subtle in nature, yet these can be harnessed with the use of neurophysiological tools. Tools like Eye-gaze tracking and Electroencephalography (EEG) offer avenues to infer attention, memory load, and decision-making. Accessing such neurophysiological signals need one to deal with technical barriers, e.g., multi-modal synchronization, latency constraints, and real-time signal processing, hindering their adoption in dynamic learning environments. Here, we present the architecture that can be used to realize and overcome the technological challenges faced while integrating the cueing paradigm with synchronized multi-modal neurophysiological signal acquisition. This platform (i) estimates cognitive load through synchronized eye-tracking and EEG data during task execution, (ii) correlates it with performance outcomes, and (iii) generates adaptive cues tailored to individual cognitive profiles to optimize learning efficiency. Further, by integrating with existing platforms like LAReflect, our approach provides actionable feedback for both learners and trainers. The broad aim is to enable implementation of cognition-aware skill learning platforms with adaptive, individualized cueing to foster effective learning.

This study presents a computational investigation of a Rotating Detonation Engine (RDE) integrated with Tesla valve-based injectors to mitigate backpressure-induced by the detonation wave. The RDE configuration includes an inlet, a plenum, and a deflagration-to-detonation combustion chamber, with and without Tesla valves to compare the regulation of the inlet flow. The base RDE suffers from reverse flow of high-temperature gases, posing flashback risks and unstable operation. A two-dimensional unwrapped RDE model is simulated using the transient Navier–Stokes equations coupled with energy and species transport equations, a detailed chemical kinetics model, and the ideal gas equation of state. Results reveal that the inclusion of Tesla valves effectively suppresses reverse flow, stabilizes inlet conditions, and maintains robust detonation wave propagation. Compared to the base case, the Tesla valve-enhanced RDE exhibits reduced pressure fluctuations, narrowed from 2.8–6.4 atm to 3.2–5.5 atm. These findings highlight the potential of Tesla valve configurations for improving operational stability in RDEs. Future work will focus on geometric optimization and experimental validation.

Soil-cement bricks are vastly more energy efficient than fired clay bricks. Although they have adequate strength, they absorb high levels of moisture and in humid conditions they become soft and non-uniform expansion leads to excessive deformation and cracking. In this research a barrier layer on their surface that impedes moisture ingress is developed by depositing calcite using bacteria. Soil-cement bricks (230 mm x 110 mm x 60-75 mm) were prepared by mixing bacteria (Bacillus megaterium) and cured by spraying a nutrient media for 28 days. The specimens were tested for water absorption, porosity and compressive strength and compared with control specimens. Results showed that the rate of water absorption and porosity were significantly reduced and compressive strength was enhanced in bacteria treated bricks. The results suggest that the barrier layer created by bacterial activity greatly alleviates the weaknesses of energy efficient soil-cement bricks enabling their large scale use.

We investigate the automatic generation of Wikipedia articles as an alternative to its manual creation. We propose a framework for creating a Wikipedia article for a named entity which not only looks similar to other Wikipedia articles in its category but also aggregates ihe diverse aspects related to that named entity from the Web. In particular, a semi-supervised method is used for determining the headings and identifying the content for each heading in the Wikipedia article generated. Evaluations show that articles created by our system for categories like actors are more reliable and informative compared to those generated by previous approaches of Wikipedia article automation.

The Stanford University Unstructured design is a tool for solving multi-disciplinary problems governed by the Partial Differential Equations on general, unstructured grids. The Overset algorithm ‘Overset Parallel Engine for AeRodynamic Applications’ developed at Institute of High Performance Computing, Singapore allows dynamic overset grid capabilities with any number of bodies using parallel processing. This paper presents the coupling of OPERA with SU2 for complex multi- body aerospace problems. The modified software suite is tested for the flow past a three dimensional sphere. The results compared with analytical results support the potential and working of the algorithm compared to a non-overlapping grid. The same problem is analyzed for moving overlapping grid. The velocity contours at different time instance are shown and are found to be continuous across the overlapping grid.

Carbon Capture and Storage (CCS) is crucial for reducing emissions in hard-to-decarbonize sectors such as cement, steel, and power generation. For India, heavily reliant on coal, CCS provides a means to lower emissions while maintaining energy security. While global CCS deployment has gained momentum through policy support and funding, high costs remain a significant challenge. In India, basalt formations, especially in the Deccan Traps, present a promising and cost-effective option for CO2 storage through mineralization. India's CCS policy is still in development, with short-term incentives like carbon credits and potential long-term goals, such as carbon taxes, being considered. To scale CCS, India requires robust funding mechanisms, technological innovation, and large-scale CCS clusters. International collaboration and investment will be essential in addressing financial and technological barriers. With the right infrastructure, research, and policy framework, CCS can become a key component of India's strategy to meet its net-zero target by 2070.