Search

The impact of air pollution on the roadside plants in the Hadapsar suburban was determined using six different plant species. To assess the effects of air pollutants on roadside plants of Hadapsar area in Pune city, APTI (Air Pollution Tolerance Index) was determined of six plant species. Plants species selected for the study were Polyalthia long folia, Ficus racemosa, Terminalia catappa, Ficus benghalensis, Ficus religiosa, Samanea saman of nearly the same age size and were found in abundance. Physiological and biochemical parameters such as pH, Relative water content (RWC), Total Chlorophyll (T.C.), Ascorbic acid (A.A.) contents were determined to calculate APTI. Results indicated that all of the plants selected for the study were found to be sensitive & heavily impacted due to air pollution. Thus, these plants can be used as Bio-indicators in monitoring air pollution.

Existing studies show that voters not only dislike violence but also sanction parties and candidates for using it. But how do we reconcile these findings with the fairly frequent incidence of political violence in democracies? Why do voters in contexts with free and fair elections, mature parties, and high electoral competition vote for violent politicians? We argue that the scope for sanctioning effects is more constrained than commonly assumed, being limited to non-supporters of the violent party. Core and swing voters, on the other hand, can be mobilized through violence rather than being put off by it, leading sanctioning to fail. We expect that voters are less likely to sanction violence when it is presented as a response to a threat or injustice. Moreover, when presented with a narrative justifying violence, we anticipate that voters not only support violence, but may even sanction peaceful responses by the violent party.

We will examine three sets of questions. The first pertains to the propensity to believe in political misinformation. Specifically, we will examine if citizens are more (less) likely to believe in misinformation alleging favoritism towards outgroups (ingroups); and if they are more likely to believe in misinformation spread by co-partisans. The second question relates the efficacy of correction of political misinformation. Particularly, we study the conditions under which correction of misinformation is works. Third, we examine if misinformation leads to greater distrust and polarization among all citizens; or only among citizens with strong prejudices against outgroups.

Social media such as Whatsapp and Facebook have become important political tools in India and across the world, but their use has potentially worrisome consequences. Political elites use social media to mobilize citizens on the basis of partisan, religious, or ethnic cleavages, but preexisting hostilities associated with these cleavages lend themselves to misinterpretation and manipulation. Misinformation shared on social media could increase polarisation, amplify existing biases against other groups, or even contribute to violence. Our project explores citizens' beliefs in misinformation on social media and their consequences for attitudes. This document describes a pre-analysis plan for a study on misinformation based on a priming and vignette experiment. In the vignette experiment (experiment 1), we construct a vignette that mimics a misinformation message common in the Indian context. We examine whether corrective information is effective and whether effectiveness is conditional on respondent religious and partisan identities. In the priming experiment (experiment 2), we examine downstream consequences of priming respondents with misinformation messages. To that end, the entire vignette setup itself is considered a treatment. We test how the messages affect respondents' attitudes towards in-and out-groups. We explore whether these downstream effects are conditional on corrective messages and respondents' partisanship. The study is pre-registered prior to researcher access to outcome data and has received ethics approval from the University of Amsterdam's Ethics Review Board #2021-AISSR-13196.

Existing studies show that voters not only dislike violence but also sanction parties and candidates for using it. But how do we reconcile these findings with the fairly frequent incidence of political violence in democracies? Why do voters in contexts with free and fair elections, mature parties, and high electoral competition vote for violent politicians? We argue that the scope for sanctioning effects is more constrained than commonly assumed, being limited to non-supporters of the violent party. Core and swing voters, on the other hand, can be mobilized through violence rather than being put off by it, leading sanctioning to fail. We expect that voters are less likely to sanction violence when it is presented as a response to a threat or injustice. Moreover, when presented with a narrative justifying violence, we anticipate that voters not only support violence, but may even sanction peaceful responses by the violent party. Empirically, the paper relies on vignette experiments embedded in a representative survey of registered voters. We field our survey during elections in Uttar Pradesh, a state of more than 200 million in India, and a context that fits our scope conditions of free and fair elections while also experiencing persistent low-intensity violence against religious minorities.

Deep Convolutional Neural Networks (DCNNs) have demonstrated impressive robustness to recognize objects under transformations (eg. blur or noise) when these transformations are included in the training set. A hypothesis to explain such robustness is that DCNNs develop invariant neural representations that remain unaltered when the image is transformed. However, to what extent this hypothesis holds true is an outstanding question, as robustness to transformations could be achieved with properties different from invariance, eg. parts of the network could be specialized to recognize either transformed or non-transformed images. This paper investigates the conditions under which invariant neural representations emerge by leveraging that they facilitate robustness to transformations beyond the training distribution. Concretely, we analyze a training paradigm in which only some object categories are seen transformed during training and evaluate whether the DCNN is robust to transformations across categories not seen transformed. Our results with state-of-the-art DCNNs indicate that invariant neural representations do not always drive robustness to transformations, as networks show robustness for categories seen transformed during training even in the absence of invariant neural representations. Invariance only emerges as the number of transformed categories in the training set is increased. This phenomenon is much more prominent with local transformations such as blurring and high-pass filtering than geometric transformations such as rotation and thinning, which entail changes in the spatial arrangement of the object. Our results contribute to a better understanding of invariant neural representations in deep learning and the conditions under which it spontaneously emerges.