Search Results

Now showing 1 - 10 of 32
  • Item
    This looks More Like that: Enhancing Self-Explaining Models by prototypical relevance propagation: This Looks More Like That
    (Amsterdam : Elsevier, 2022) Gautam, Srishti; Höhne, Marina M.-C.; Hansen, Stine; Jenssen, Robert; Kampffmeyer, Michael
    Current machine learning models have shown high efficiency in solving a wide variety of real-world problems. However, their black box character poses a major challenge for the comprehensibility and traceability of the underlying decision-making strategies. As a remedy, numerous post-hoc and self-explanation methods have been developed to interpret the models’ behavior. Those methods, in addition, enable the identification of artifacts that, inherent in the training data, can be erroneously learned by the model as class-relevant features. In this work, we provide a detailed case study of a representative for the state-of-the-art self-explaining network, ProtoPNet, in the presence of a spectrum of artifacts. Accordingly, we identify the main drawbacks of ProtoPNet, especially its coarse and spatially imprecise explanations. We address these limitations by introducing Prototypical Relevance Propagation (PRP), a novel method for generating more precise model-aware explanations. Furthermore, in order to obtain a clean, artifact-free dataset, we propose to use multi-view clustering strategies for segregating the artifact images using the PRP explanations, thereby suppressing the potential artifact learning in the models.
  • Item
    Regional income and wave energy deployment in Ireland
    (Oxford : Wiley-Blackwell, 2020) Farrell, Niall; O'Donoghue, Cathal; Morrissey, Karyn
    Alongside environmental benefits, renewable energy deployment is often evaluated on grounds of regional development. Focusing on wave energy deployment in Ireland, this paper quantifies employment-related welfare change net of associated subsidy costs. Although the added employment reduces inter-regional inequality, certain subsidies increase total income inequality by a greater extent. Total inequality increases by 0.25% in the preferred scenario. This pattern of incidence persists under an optimistic scenario where all manufacturing activity is carried out locally. This finding highlights that policies of regional development should consider the spatial distribution of associated subsidy costs.
  • Item
    European H2020 Project WORTECS Wireless Mixed Reality Prototyping
    (Oulu : Academy Publisher, 2020) Bouchet, Olivier; O'Brien, Dominic; Singh, Ravinder; Faulkner, Grahame; Ghoraishi, Mir; Garcia-Marquez, Jorge; Vercasson, Guillaume; Brzozowski, Marcin; Sark, Vladica
    This paper presents European collaborative project WORTECS objectives and reports on the development of several radio and optical wireless prototypes and a demonstrator targeting mixed reality (MR) application. The aim is to achieve a net throughput of up to Tbps in an indoor heterogeneous network for the MR use case, which seems to be a high throughput "killer application" beyond 5G. A special routing device is associated with the demonstrator to select the most suitable wireless access technology. Post introduction to the project, an overview of the demonstrator is presented with details of the current progress of the prototypes.
  • Item
    Reply to Ruhl and Craig: Assessing and governing extreme climate risks needs to be legitimate and democratic
    (Washington, DC : National Acad. of Sciences, 2022) Kemp, Luke; Xu, Chi; Depledge, Joanna; Ebi, Kristie L.; Gibbins, Goodwin; Kohler, Timothy A.; Rockström, Johan; Scheffer, Marten; Schellnhuber, Hans Joachim; Steffen, Will; Lenton, Timothy M.
    [No abstract available]
  • Item
    Reply to Burgess et al: Catastrophic climate risks are neglected, plausible, and safe to study
    (Washington, DC : National Acad. of Sciences, 2022) Kemp, Luke; Xu, Chi; Depledge, Joanna; Ebi, Kristie L.; Gibbins, Goodwin; Kohler, Timothy A.; Rockström, Johan; Scheffer, Marten; Schellnhuber, Hans Joachim; Steffen, Will; Lenton, Timothy M.
  • Item
    Gas plasma–oxidized sodium chloride acts via hydrogen peroxide in a model of peritoneal carcinomatosis
    (Washington, DC : National Acad. of Sciences, 2022) Miebach, Lea; Freund, Eric; Clemen, Ramona; Kersting, Stephan; Partecke, Lars-Ivo; Bekeschus, Sander
    Gas plasma technology generates reactive oxygen and nitrogen species (ROS/RNS), inducing lethal oxidative damage in tumor cells. The transfer of gas plasma–derived ROS/RNS into liquids has been proposed as an innovative anti-cancer strategy targeting peritoneal carcinomatosis (PC). However, the mechanism of action is under debate. To this end, we compared gas plasma–oxidized medical-grade sodium chloride (oxNaCl) with a concentration-matched control (cmc) of NaCl enriched with equivalent concentrations of H2O2 and NO32 in several cell lines and models of PC. Strikingly, oxNaCl and cmc performed equally well in oxidation and cytotoxic activity in tumor cells in two-dimensional cultures, three-dimensional (3D) tumor spheroids, vascularized 3D tumors grown on chicken-embryo chorioallantoic membranes, and a syngeneic PC mouse model in vivo. Given the importance of immunotherapies in oncology today, we focused on immunological consequences of the treatment. Again, to a similar extent, oxNaCl and cmc increased tumor cell immunogenicity and enhanced uptake by and maturation of peripheral blood monocyte–derived dendritic cells together with an inflammatory secretion profile. Furthermore, NanoString gene expression profiling revealed immune system processes and unfolded protein response-related pathways as being linked to the observed anti-tumor effects for both oxNaCl and cmc. In conclusion, gas plasma–generated oxNaCl and cmc showed equal therapeutic efficacy in our PC-related models. In light of the many promising anti-cancer studies of gas plasma–oxidized liquids and the convenient production of corresponding cmcs in large quantities as needed in clinics, our findings may spur research lines based on low-dose oxidants in peritoneal cancer therapy.
  • Item
    Trajectories of the Earth System in the Anthropocene
    (Washington, DC : NAS, 2018) Steffen, Will; Rockström, Johan; Richardson, Katherine; Lenton, Timothy M.; Folke, Carl; Liverman, Diana; Summerhayes, Colin P.; Barnosky, Anthony D.; Cornell, Sarah E.; Crucifix, Michel; Donges, Jonathan F.; Fetzer, Ingo; Lade, Steven J.; Scheffer, Marten; Winkelmann, Ricarda; Schellnhuber, Hans Joachim
    We explore the risk that self-reinforcing feedbacks could push the Earth System toward a planetary threshold that, if crossed, could prevent stabilization of the climate at intermediate temperature rises and cause continued warming on a “Hothouse Earth” pathway even as human emissions are reduced. Crossing the threshold would lead to a much higher global average temperature than any interglacial in the past 1.2 million years and to sea levels significantly higher than at any time in the Holocene. We examine the evidence that such a threshold might exist and where it might be. If the threshold is crossed, the resulting trajectory would likely cause serious disruptions to ecosystems, society, and economies. Collective human action is required to steer the Earth System away from a potential threshold and stabilize it in a habitable interglacial-like state. Such action entails stewardship of the entire Earth System—biosphere, climate, and societies—and could include decarbonization of the global economy, enhancement of biosphere carbon sinks, behavioral changes, technological innovations, new governance arrangements, and transformed social values.
  • Item
    Stewardship of global collective behavior
    (Washington, DC : National Acad. of Sciences, 2021) Bak-Coleman, Joseph B.; Alfano, Mark; Barfuss, Wolfram; Bergstrom, Carl T.; Centeno, Miguel A.; Couzin, Iain D.; Donges, Jonathan F.; Galesic, Mirta; Gersick, Andrew S.; Jacquet, Jennifer; Kao, Albert B.; Moran, Rachel E.; Romanczuk, Pawel; Rubenstein, Daniel I.; Tombak, Kaia J.; Van Bavel, Jay J.; Weber, Elke U.
    Collective behavior provides a framework for understanding how the actions and properties of groups emerge from the way individuals generate and share information. In humans, information flows were initially shaped by natural selection yet are increasingly structured by emerging communication technologies. Our larger, more complex social networks now transfer high-fidelity information over vast distances at low cost. The digital age and the rise of social media have accelerated changes to our social systems, with poorly understood functional consequences. This gap in our knowledge represents a principal challenge to scientific progress, democracy, and actions to address global crises. We argue that the study of collective behavior must rise to a “crisis discipline” just as medicine, conservation, and climate science have, with a focus on providing actionable insight to policymakers and regulators for the stewardship of social systems.
  • Item
    Social tipping dynamics for stabilizing Earth's climate by 2050
    (2020) Otto, Ilona M.; Donges, Jonathan F.; Cremades, Roger; Bhowmik, Avit; Hewitt, Richard J.; Lucht, Wolfgang; Rockström, Johan; Allerberger, Franziska; McCaffrey, Mark; Doe, Sylvanus S.P.; Lenferna, Alex; Morán, Nerea; van Vuuren, Detlef P.; Schellnhuber, Hans Joachim
    Safely achieving the goals of the Paris Climate Agreement requires a worldwide transformation to carbon-neutral societies within the next 30 y. Accelerated technological progress and policy implementations are required to deliver emissions reductions at rates sufficiently fast to avoid crossing dangerous tipping points in the Earth's climate system. Here, we discuss and evaluate the potential of social tipping interventions (STIs) that can activate contagious processes of rapidly spreading technologies, behaviors, social norms, and structural reorganization within their functional domains that we refer to as social tipping elements (STEs). STEs are subdomains of the planetary socioeconomic system where the required disruptive change may take place and lead to a sufficiently fast reduction in anthropogenic greenhouse gas emissions. The results are based on online expert elicitation, a subsequent expert workshop, and a literature review. The STIs that could trigger the tipping of STE subsystems include 1) removing fossil-fuel subsidies and incentivizing decentralized energy generation (STE1, energy production and storage systems), 2) building carbon-neutral cities (STE2, human settlements), 3) divesting from assets linked to fossil fuels (STE3, financial markets), 4) revealing the moral implications of fossil fuels (STE4, norms and value systems), 5) strengthening climate education and engagement (STE5, education system), and 6) disclosing information on greenhouse gas emissions (STE6, information feedbacks). Our research reveals important areas of focus for larger-scale empirical and modeling efforts to better understand the potentials of harnessing social tipping dynamics for climate change mitigation.
  • Item
    Communicating sentiment and outlook reverses inaction against collective risks
    (Washington, DC : National Acad. of Sciences, 2020) Wang, Zhen; Jusup, Marko; Guo, Hao; Shi, Lei; GeÄŤek, SunÄŤana; Anand, Madhur; Perc, MatjaĹľ; Bauch, Chris T.; Kurths, JĂĽrgen; Boccaletti, Stefano; Schellnhuber, Hans Joachim
    Collective risks permeate society, triggering social dilemmas in which working toward a common goal is impeded by selfish interests. One such dilemma is mitigating runaway climate change. To study the social aspects of climate-change mitigation, we organized an experimental game and asked volunteer groups of three different sizes to invest toward a common mitigation goal. If investments reached a preset target, volunteers would avoid all consequences and convert their remaining capital into monetary payouts. In the opposite case, however, volunteers would lose all their capital with 50% probability. The dilemma was, therefore, whether to invest one's own capital or wait for others to step in. We find that communicating sentiment and outlook helps to resolve the dilemma by a fundamental shift in investment patterns. Groups in which communication is allowed invest persistently and hardly ever give up, even when their current investment deficits are substantial. The improved investment patterns are robust to group size, although larger groups are harder to coordinate, as evidenced by their overall lower success frequencies. A clustering algorithm reveals three behavioral types and shows that communication reduces the abundance of the free-riding type. Climate-change mitigation, however, is achieved mainly by cooperator and altruist types stepping up and increasing contributions as the failure looms. Meanwhile, contributions from free riders remain flat throughout the game. This reveals that the mechanisms behind avoiding collective risks depend on an interaction between behavioral type, communication, and timing.