Rich Space Research Articles

Image Captioning via Dynamic Path Customization.

This article explores a novel dynamic network for vision and language (V&L) tasks, where the inferring structure is customized on the fly for different inputs. Most previous state-of-the-art (SOTA) approaches are static and handcrafted networks, which not only heavily rely on expert knowledge but also ignore the semantic diversity of input samples, therefore resulting in suboptimal performance. To address these issues, we propose a novel Dynamic Transformer Network (DTNet) for image captioning, which dynamically assigns customized paths to different samples, leading to discriminative yet accurate captions. Specifically, to build a rich routing space and improve routing efficiency, we introduce five types of basic cells and group them into two separate routing spaces according to their operating domains, i.e., spatial and channel. Then, we design a Spatial-Channel Joint Router (SCJR), which endows the model with the capability of path customization based on both spatial and channel information of the input sample. To validate the effectiveness of our proposed DTNet, we conduct extensive experiments on the MS-COCO dataset and achieve new SOTA performance on both the Karpathy split and the online test server. The source code is publicly available at https://github.com/xmu-xiaoma666/DTNet.

"Wise Mothers," "Mom Bugs," and Pyŏngmat (Twisted Tastes): The Limits of Maternal Emotional Expression in South Korean Webtoons

Abstract: In the last decade, South Korean webtoons have joined television dramas (K-dramas) and K-pop as an important element of the Korean Wave abroad. Domestically, the South Korean public can idly browse or religiously follow thousands of free or subscription webtoons on their smartphones. Webtoon artists may dream of achieving broader success by having their works adapted transmedially, as a web drama (online TV drama) or even better, as a network, cable or Netflix series that achieves mainstream success. Two significant subgenres of webtoons are "lifestyle webtoons" (saenghwarwept'un) and "family webtoons" (kajogwept'un) and these two subgenres are combined in what I am calling "childcare webtoons" (yugawept'un). While South Korea also follows global media trends of "mom blogs" and Instagram microcelebrities or Youtubers sharing their parenting journeys ("sharenting"), "childcare webtoons" are a particularly rich space for observing changing mothering ideologies and parenting norms. This article examines "mom humor" and other stories mothers tell across the South Korean Internet, paying particular attention to what kinds of emotional expression are sanctioned and what is taboo. I analyze webtoons such as "I'm a Mom (Nanŭn ŏmmada)" and "The Birth of a Married Woman(Yubunyŏŭi t'ansaeng)," suggesting that even as the tedium of everyday motherhood is increasingly critiqued through media, the still somewhat rigid gender roles and maternal expectations in South Korea translate into an overall more muted and subtle "mom humor" alongside the still more socially expected stories of maternal gratitude and fulfillment.

ДІАСПОРА ЯК ГЕТЕРОТОПІЯ В РОМАНІ МАЙКЛА ОНДАТЖЕ “У ШКІРІ ЛЕВА”

The present article examines the postcolonial novel “In the Skin of a Lion” written by Michael Ondaatje, one of Canada’s leading contemporary writers. As the marginalized experience of an immigrant had a signifi cant impact on shaping the author’s diasporic identity, Ondaatje, originally from Ceylon and a Canadian citizen, depicts in his novel the immigrant experience and the distinctiveness of the corresponding diasporic space within a linguistic and cultural community. The article underscores that the space, which both refl ects and distorts the real world, can be characterized as a heterotopian one. In multicultural societies, diaspora serves as one such space – an entity that unites individuals bound by their shared ethnic heritage and brings together inhabitants who often fi nd themselves isolated and marginalized. In the novel “In the Skin of a Lion”, the immigrant district is depicted as a space with utopian features. A cornerstone of the analysis lies in the observation of how immigrants navigate and interact with foreign landscapes. The article demonstrates how these individuals utilize the canvas of a foreign country to reimagine and reconstruct familiar environments, a process that both mirrors and refracts their original contexts. Signifi cantly, the diasporic space inhabited by the characters is similar to Foucault’s concept of a heterotopic mirror, a realm without a concrete locus yet integral to shaping the characters’ experiences and perspectives. The methodology of the article relies primarily on the notion of heterotopia by French philosopher Michel Foucault, Homi Bhabha’s explorations and his concept of an intermediate space or the “third” space, and Edward Said’s postcolonial theory as well as other cultural and literary studies pertaining to the diaspora. It is shown that the heterotopian space in M. Ondaatje’s novel is a place for the formation of the diasporic identity of his characters. Ultimately, according to Foucault’s principles of heterotopia, in the novel “In the Skin of a Lion”, the diaspora and the immigrant community’s enclave are portrayed as a rich and transformative heterotopic space. Key words: Michael Ondaatje, “In the Skin of a Lion,” Michel Foucault, heterotopia, space, diasporic identity.

Inverse design of mechanical metamaterial achieving a prescribed constitutive curve

Besides exhibiting excellent capabilities such as energy absorption, phase-transforming metamaterials offer a vast design space for achieving nonlinear constitutive relations. This is facilitated by switching between different patterns under deformation. However, the related inverse design problem is quite challenging, due to the lack of appropriate mathematical formulation and the convergence issue in the post-buckling analysis of intermediate designs. In this work, periodic unit cells are explicitly described by the moving morphable voids method and effectively analyzed by eliminating the degrees of freedom in void regions. Furthermore, by exploring the Pareto frontiers between error and cost, an inverse design formulation is proposed for unit cells. This formulation aims to achieve a prescribed constitutive curve and is validated through numerical examples and experimental results. The design approach presented here can be extended to the inverse design of other types of mechanical metamaterials with prescribed nonlinear effective properties.

Concept Decomposition for Visual Exploration and Inspiration

A creative idea is often born from transforming, combining, and modifying ideas from existing visual examples capturing various concepts. However, one cannot simply copy the concept as a whole, and inspiration is achieved by examining certain aspects of the concept. Hence, it is often necessary to separate a concept into different aspects to provide new perspectives. In this paper, we propose a method to decompose a visual concept, represented as a set of images, into different visual aspects encoded in a hierarchical tree structure. We utilize large vision-language models and their rich latent space for concept decomposition and generation. Each node in the tree represents a sub-concept using a learned vector embedding injected into the latent space of a pretrained text-to-image model. We use a set of regularizations to guide the optimization of the embedding vectors encoded in the nodes to follow the hierarchical structure of the tree. Our method allows to explore and discover new concepts derived from the original one. The tree provides the possibility of endless visual sampling at each node, allowing the user to explore the hidden sub-concepts of the object of interest. The learned aspects in each node can be combined within and across trees to create new visual ideas, and can be used in natural language sentences to apply such aspects to new designs. Project page: https://inspirationtree.github.io/inspirationtree/

GarmentCode: Programming Parametric Sewing Patterns

Garment modeling is an essential task of the global apparel industry and a core part of digital human modeling. Realistic representation of garments with valid sewing patterns is key to their accurate digital simulation and eventual fabrication. However, little-to-no computational tools provide support for bridging the gap between high-level construction goals and low-level editing of pattern geometry, e.g., combining or switching garment elements, semantic editing, or design exploration that maintains the validity of a sewing pattern. We suggest the first DSL for garment modeling - GarmentCode - that applies principles of object-oriented programming to garment construction and allows designing sewing patterns in a hierarchical, component-oriented manner. The programming-based paradigm naturally provides unique advantages of component abstraction, algorithmic manipulation, and free-form design parametrization. We additionally support the construction process by automating typical low-level tasks like placing a dart at a desired location. In our prototype garment configurator, users can manipulate meaningful design parameters and body measurements, while the construction of pattern geometry is handled by garment programs implemented with GarmentCode. Our configurator enables the free exploration of rich design spaces and the creation of garments using interchangeable, parameterized components. We showcase our approach by producing a variety of garment designs and retargeting them to different body shapes using our configurator. The library and garment configurator are available at https://github.com/maria-korosteleva/GarmentCode.

C-Shells: Deployable Gridshells with Curved Beams

We introduce a computational pipeline for simulating and designing C-shells , a new class of planar-to-spatial deployable linkage structures. A C-shell is composed of curved flexible beams connected at rotational joints that can be assembled in a stress-free planar configuration. When actuated, the elastic beams deform and the assembly deploys towards the target 3D shape. We propose two alternative computational design approaches for C-shells: (i) Forward exploration simulates the deployed shape from a planar beam layout provided by the user. Once a satisfactory overall shape is found, a subsequent design optimization adapts the beam geometry to reduce the elastic energy of the linkage while preserving the target shape. (ii) Inverse design is facilitated by a new geometric flattening method that takes a design surface as input and computes an initial layout of piecewise straight linkage beams. Our design optimization algorithm then calculates the smooth curved beams to best reproduce the target shape at minimal elastic energy. We find that C-shells offer a rich space for design and show several studies that highlight new shape topologies that cannot be achieved with existing deployable linkage structures.

Harvesting Twitter Data for Studying Motor Behavior in Disabled Populations: An Introduction and Tutorial in Python.

Social media platforms are rich and dynamic spaces where individuals communicate on a person-to-person level and to broader audiences. These platforms provide a wealth of publicly available data that can shed light on the lived experiences of people from numerous clinical populations. Twitter can be used to examine individual expressions and community discussions about specific characteristics (e.g., motor skills, burnout) associated with a diagnostic group. These data are useful for understanding the perspectives of a diverse, international group of self-advocates representing a wide range of clinical populations. Here, we provide a framework for how to harvest data from Twitter through their free, academic researcher application programming interface access using Python, a free, open-source programming language. We also provide a sample data set harvested using this framework and a set of analyses on these data specifically related to motor differences in neurodevelopmental conditions. This framework offers a cost-effective and flexible means of harvesting and analyzing Twitter data. Researchers should utilize these resources to advance our understanding of the lived experiences of clinical populations through social media platforms and to determine the critical questions that are of most importance to improving quality of life.

Peace education in a time of war: the Museum of Peace in Rivne, Ukraine as a space of memory making and hope

ABSTRACT Peace museums play an important role in peace education by offering visitors informal and non-formal education. As sites of remembrance, peace museums are rich pedagogical spaces for experiential learning and reflection. Educating children in the spirit of peace, tolerance and harmony between nations has been central to the work of the Museum of Peace in Rivne in Ukraine. Whilst peace museums usually engage in peacebuilding and memory making in times of peace, post conflict, this article reports on the work of the Museum in Rivne in a time of war. Wartime brings difficult questions about engaging in peacebuilding in the face of military aggression and about sustaining memory-making work when violent conflict makes memories too immediate and painful. As explained in the article, the reinvigorated peacebuilding effort at the Museum in Rivne demonstrates that, in a time of war, it is even more important to promote peace, in opposition to war. Through the annual event ‘I Vote for Peace’, the Museum has sought to create a network of schools committed to global tolerance and peacekeeping, as well as offer Ukrainian children a space for talking about their experiences and their hope for a peaceful future.

Snowmass2021 cosmic frontier white paper: Ultraheavy particle dark matter

We outline the unique opportunities and challenges in the search for “ultraheavy” dark matter candidates with masses between roughly 10 TeV and the Planck scale m_{\rm pl} ≈ 10^{16} TeV. This mass range presents a wide and relatively unexplored dark matter parameter space, with a rich space of possible models and cosmic histories. We emphasize that both current detectors and new, targeted search techniques, via both direct and indirect detection, are poised to contribute to searches for ultraheavy particle dark matter in the coming decade. We highlight the need for new developments in this space, including new analyses of current and imminent direct and indirect experiments targeting ultraheavy dark matter and development of new, ultra-sensitive detector technologies like next-generation liquid noble detectors, neutrino experiments, and specialized quantum sensing techniques.

Who Controls the Agenda Controls the Legislature

We model legislative decision-making with an agenda setter who can propose policies sequentially, tailoring each proposal to the status quo that prevails after prior votes. Voters are sophisticated, and the agenda setter cannot commit to future proposals. Nevertheless, the agenda setter obtains her favorite outcome in every equilibrium regardless of the initial default policy. Central to our results is a new condition on preferences, manipulability, that holds in rich policy spaces, including spatial settings and distribution problems. Our findings therefore establish that, despite the sophistication of voters and the absence of commitment power, the agenda setter is effectively a dictator. (JEL D71, D72, D78)

Fish hearing revealed: Do we understand hearing in critical fishes and marine tetrapods.

Hearing evolved in lampreys with a frequency range of 50-200 Hz. This hearing range is comparable to that of elasmobranchs, most non-teleosts, and lungfish. Elasmobranchs most likely use the saccule and the papilla neglecta (PN) for hearing. In non-teleosts and teleosts, lungfish, and certain tetrapods the saccule is the likely sensor for sound reception while the lagena and the PN are important for gravistatic sensing. Coelacanth and most tetrapods have a basilar papilla (BP) for hearing. In coelacanth and tetrapods, the hair cells of the BP are in contact with a basilar and a tectorial membrane. These membranes transmit mechanical vibrations. A cochlear aqueduct (CA) provides a connection between the cerebrospinal fluid that has a sodium rich space in coelacanth and tetrapods while the potassium rich endolymph is known in vertebrates. A unique feature is known in basic sarcopterygians, the intracranial joint, that never developed in actinopterygians and has been lost in lungfish and tetrapods. The BP in coelacanths is thought to generate pressure with the intracranial joint that will be transmitted to the CA. Lungs or a swim bladder are not forming in Chondrichthyes, structures that have a major impact on hearing in teleosts and tetrapods.

Resident Population Density-Inspired Deployment of K-Tier Aerial Cellular Network

Using Unmanned Aerial Vehicles (UAVs) to enhance network coverage has proven a variety of benefits compared to terrestrial counterparts. One of the commonly used mathematical tools to model the locations of the UAVs is stochastic geometry (SG). However, in the existing studies, both users and UAVs are often modeled as homogeneous point processes. In this paper, we consider an inhomogeneous Poisson point process (PPP)-based model for the locations of the users that captures the degradation in the density of active users as we move away from the town center. In addition, we propose the deployment of aerial vehicles following the same inhomogeneity of the users to maximize the performance. In addition, a multi-tier network model is also considered to make better use of the rich space resources. Then, the analytical expressions of the coverage probability for a typical user and the total coverage probability are derived. Finally, we optimize the coverage probability with limitations of the total number of UAVs and the minimum local coverage probability. Finally we give the optimal UAV distribution parameters when the maximum overall coverage probability is reached.

A 3D Lithiophilic Host for Dendrite-Free Lithium Metal Anode via One-Step Carbonization of an Energetic Metal-Organic Framework.

Low Coulombic efficiency (CE) and safety issues are huge problems that hinder the practical application of Li metal anodes. Constructing Li host structures decorated with functional species can restrain the growth of Li dendrites and alleviate the great volume change. Here, a 3D porous carbonaceous skeleton modified with rich lithiophilic groups (Zn, ZnO, and Zn(CN)2 ) is synthesized as a Li host via one-step carbonization of a triazole-containing metal-organic framework. The nano lithiophilic groups serve as preferred sites for Li nucleation and growth, regulating a uniform Li+ flux and uniform current density distribution. In addition, the 3D porous network functions as a Li reservoir that provides rich internal space to store Li, thus alleviating the volumetric expansion during Li plating/stripping process. Thanks to these component and structural merits, an ultra-low overpotential for Li deposition is achieved, together with high CE of over 99.5% for more than 500 cycles at 1mAcm-2 and 1mAhcm-2 in half cells. The symmetric cells exhibit a prolonged cycling of 900h at 1mAcm-2 . The full cells by coupling Zn/ZnO/Zn(CN)2 @C-Li anode with LiFePO4 cathode deliver a high capacity retention of 94.3% after 200 cycles at 1C.

Conformally graded metamaterials for elastic wave guidance

Although metamaterials have been widely used for controlling elastic waves through bandgap engineering, the directed guidance of stress waves in non-periodic structures has remained a challenge. This work demonstrates that spatially graded metamaterials based on conformal mappings present a rich design space for controlling and attenuating wave motion — without the need for bandgaps. Conformal mappings transform an elementary unit cell by scaling and rotation into graded lattices with approximately geometrically similar unit cells. This self-similarity allows for control over the local wave dispersion throughout the metamaterial. As a key mechanism, it is shown that elastic waves cannot propagate through graded unit cells with significant size differences, except at low frequencies. This is exploited to create low-pass elastic wave guides, extending beyond classical bandgap engineering, since bandgaps are not required to achieve wave guiding and attenuation. Experiments confirm the low-pass elastic wave filtering capability of a planar truss metamaterial with conformal grading. Finally, a systematic design of curved metamaterial surfaces is presented, providing a flexible framework for programming low-pass attenuation and wave guiding in three dimensions.

Selective Explanations: Leveraging Human Input to Align Explainable AI

While a vast collection of explainable AI (XAI) algorithms has been developed in recent years, they have been criticized for significant gaps with how humans produce and consume explanations. As a result, current XAI techniques are often found to be hard to use and lack effectiveness. In this work, we attempt to close these gaps by making AI explanations selective ---a fundamental property of human explanations---by selectively presenting a subset of model reasoning based on what aligns with the recipient's preferences. We propose a general framework for generating selective explanations by leveraging human input on a small dataset. This framework opens up a rich design space that accounts for different selectivity goals, types of input, and more. As a showcase, we use a decision-support task to explore selective explanations based on what the decision-maker would consider relevant to the decision task. We conducted two experimental studies to examine three paradigms based on our proposed framework: in Study 1, we ask the participants to provide critique-based or open-ended input to generate selective explanations (self-input). In Study 2, we show the participants selective explanations based on input from a panel of similar users (annotator input). Our experiments demonstrate the promise of selective explanations in reducing over-reliance on AI and improving collaborative decision making and subjective perceptions of the AI system, but also paint a nuanced picture that attributes some of these positive effects to the opportunity to provide one's own input to augment AI explanations. Overall, our work proposes a novel XAI framework inspired by human communication behaviors and demonstrates its potential to encourage future work to make AI explanations more human-compatible.

(Invited) Understanding Photoelectrocatalysis on Epitaxial Oxide Surfaces

The intermittent nature of renewable energy sources requires a clean, scalable means of converting and storing energy. One Earth abundant storage option is water electrolysis, storing energy in the bonds of O2 and H2. Photoelectrochemical (PEC) cells based on semiconductor/liquid interfaces can convert sunlight to chemical fuels without external circuitry, such as “splitting” water into O2 and H2 upon illumination. The efficacy of conversion depends in part on the location of the semiconductor band edges, dictating absorption, and also the rectifying properties of semiconductor–electrolyte junctions, which drives the separation of electron–hole pairs.We will present studies of model oxide La(1-x)SrxFeO3 photoelectrodes grown by molecular beam epitaxy (MBE) and pulsed laser deposition (PLD) that display a known crystallographic orientation, surface area, path for charge transport, and strain. Aliovalent doping can tune the transition metal redox properties as well as the material band gap and conductivity, resulting in a rich phase space of possible water-oxidation photoanodes. Comparing the photooxidation of a fast redox couple to the slugging water oxidation reaction demonstrates the potential of electronic states that must be filled prior to oxygen evolution. Measurements of the surface speciation in a humid environment demonstrate that Sr incorporation also promotes hydroxylation, suggesting that water oxidation proceeds through the oxidation of FeIII-OH states, which is facilitated by Sr incorporation and results in an increased photovoltage. Surface-sensitive measurement of the valence band and unoccupied states (via O K-edge X-ray absorption spectroscopy) under humid conditions further support this picture. This fundamental insight builds understanding necessary for the design of active, earth-abundant photocatalysts that can be integrated into PEC devices for efficient conversion of solar energy into chemical fuels.

Physiological adaptation in flagellar architecture improves Vibrio alginolyticus chemotaxis in complex environments

Bacteria navigate natural habitats with a wide range of mechanical properties, from the ocean to the digestive tract and soil, by rotating helical flagella like propellers. Species differ in the number, position, and shape of their flagella, but the adaptive value of these flagellar architectures is unclear. Many species traverse multiple types of environments, such as pathogens inside and outside a host. We investigate the hypothesis that flagellar architectures mediate environment-specific benefits in the marine pathogen Vibrio alginolyticus which exhibits physiological adaptation to the mechanical environment. In addition to its single polar flagellum, the bacterium produces lateral flagella in environments that differ mechanically from water. These are known to facilitate surface motility and attachment. We use high-throughput 3D bacterial tracking to quantify chemotactic performance of both flagellar architectures in three archetypes of mechanical environments relevant to the bacterium's native habitats: water, polymer solutions, and hydrogels. We reveal that lateral flagella impede chemotaxis in water by lowering the swimming speed but improve chemotaxis in both types of complex environments. Statistical trajectory analysis reveals two distinct underlying behavioral mechanisms: In viscous solutions of the polymer PVP K90, lateral flagella increase the swimming speed. In agar hydrogels, lateral flagella improve overall chemotactic performance, despite lowering the swimming speed, by preventing trapping in pores. Our findings show that lateral flagella are multipurpose tools with a wide range of applications beyond surfaces. They implicate flagellar architecture as a mediator of environment-specific benefits and point to a rich space of bacterial navigation behaviors in complex environments.

Controlled Assembly of CdSe Nanoplatelet Thin Films and Nanowires.

We assemble semiconductor CdSe nanoplatelets (NPs) at the air/liquid interface into 2D monolayers several micrometers wide, distinctly displaying nematic order. We show that this configuration is the most favorable energetically and that the edge-to-edge distance between neighboring NPs can be tuned by ligand exchange without disrupting film topology and nanoparticle orientation. We explore the rich assembly phase space by using depletion interactions to direct the formation of 1D nanowires from stacks of NPs. The improved control and understanding of the assembly of semiconductor NPs offers opportunities for the development of cheaper optoelectronic devices that rely on 1D or 2D charge delocalization throughout the assembled monolayers and nanowires.

Kinesin and myosin motors compete to drive rich multiphase dynamics in programmable cytoskeletal composites.

The cellular cytoskeleton relies on diverse populations of motors, filaments, and binding proteins acting in concert to enable nonequilibrium processes ranging from mitosis to chemotaxis. The cytoskeleton's versatile reconfigurability, programmed by interactions between its constituents, makes it a foundational active matter platform. However, current active matter endeavors are limited largely to single force-generating components acting on a single substrate-far from the composite cytoskeleton in cells. Here, we engineer actin-microtubule (MT) composites, driven by kinesin and myosin motors and tuned by crosslinkers, to ballistically restructure and flow with speeds that span three orders of magnitude depending on the composite formulation and time relative to the onset of motor activity. Differential dynamic microscopy analyses reveal that kinesin and myosin compete to delay the onset of acceleration and suppress discrete restructuring events, while passive crosslinking of either actin or MTs has an opposite effect. Our minimal advection-diffusion model and spatial correlation analyses correlate these dynamics to structure, with motor antagonism suppressing reconfiguration and demixing, while crosslinking enhances clustering. Despite the rich formulation space and emergent formulation-dependent structures, the nonequilibrium dynamics across all composites and timescales can be organized into three classes-slow isotropic reorientation, fast directional flow, and multimode restructuring. Moreover, our mathematical model demonstrates that diverse structural motifs can arise simply from the interplay between motor-driven advection and frictional drag. These general features of our platform facilitate applicability to other active matter systems and shed light on diverse ways that cytoskeletal components can cooperate or compete to enable wide-ranging cellular processes.