Intermodal Interactions Research Articles

Multimode Mamyshev oscillator.

We present a spatiotemporally mode-locked Mamyshev oscillator. A wide variety of multimode mode-locked states, with varying degrees of spatiotemporal coupling, are observed. We find that some control of the modal content of the output beam is possible through the cavity design. Comparison of simulations with experiments indicates that spatiotemporal mode locking (STML) is enabled by nonlinear intermodal interactions and spatial filtering, along with the Mamyshev mechanism. This work represents a first, to the best of our knowledge, exploration of STML in an oscillator with a Mamyshev saturable absorber.

Drivers' Cycling Experiences and Acceptability of Micromobility Use among Children in Ghana

Apart from their potential environmental benefits, micromobility modes of transport can afford their users enormous health benefits. This notwithstanding, their adoption is very low or non-existent in most African cities including Ghana. The paucity of research on micromobility in Ghana restricts any objective discussions on the factors that contribute to its low adoption. However, data from elsewhere flag the lack of safe spaces as an important factor. The creation of safe spaces and a micromobility supportive environment will have to be premised on the nuanced understanding of existing intermodal interactions and con flicts. Although micromobility vehicles like e-bikes and e-scooters are not common in Ghana, bicycles are, and they offer a window of opportunity to study intermodal interactions between cyclists (micromobility proxy) and other road users. Using a mixed methods approach, this study investigates the interactions and con flicts that arise between child cyclists and informal-sector commercial vehicle drivers in two Ghanaian cities (Cape Coast and Kumasi) and attempt to predict micromobility acceptability based on drivers' attitudes towards child cyclists and drivers cycling experiences. The results show a low acceptability of micromobility by drivers. They also had several negative stereotypes about cyclists and considered them illegitimate road users. Educational attainment, employment status, and priority access emerged as the best predictors of acceptability of micromobility use by informal-sector commercial drivers. Taken together, the low acceptability of micromobility and the negative perceptions drivers have of cyclists might present significant stumbling blocks to adoption of micromobility in the study areas. Further studies on the suitability of the built environment, intention-to-use, policy development as well as the development of a business model for micromobility will go a long way in shaping the next steps.

Computation of optical waveguide interaction for quantum gates implementation

The system of two coupled optical dual-mode waveguides is considered. The coupling of the system is studied to find a circuit for building a control switch for two qubit gates. The classical coupled mode theory is applied and the exact expressions for coupling coefficients are derived. The parameters of the system for performing the desired operations are numerically computed and analysed. The system describing the influence of intermodal interactions is solved numerically. The distortions are analysed.

An Initial Agent-Based Model for E-scooter Accommodation into Existing Urban Streetscapes

New forms of mobility reshape the transportation landscape, changing movement for both their users and others in the environment. The transition period during which novel forms of travel are being explored can be a challenging time while the use of spaces must be renegotiated. E-scooters, which have recently been more widely introduced to the UK, are experiencing such a moment as riders, planners, and other users of the streetscape are determining what role this technology will play in communities. The data gaps surrounding e-scooters can make this an especially difficult question for planners because of the cost of gathering relevant observational data, much of which is held under private company ownership. In light of this, this work presents an agent-based model developed to examine the integration of e-scooters into existing streetscapes. Agent-based models explore phenomena through focusing on individual behaviour and rules which in turn gives rise to emergent large scale patterns. These patterns can be dissected and interrogated with a variety of tools, allowing us to tease out individual as well as group experiences of different scenarios. An agent-based approach allows us to capture the individual behaviours of e-scooter users and those of cyclists, drivers of variously sized vehicles, pedestrians, and others present in the environment. By focusing on the interactions of these various street users, we can explore how different approaches to e-scooter integration may fare relative to varying street configurations. Their decision frameworks are informed by observational studies of e-scooter users in order to augment the available data. We discuss the current state of understanding e-scooter behaviour and the potential modelling applications, present an initial behavioural framework of e-scooter decision making and inter-modal interactions, and highlight some preliminary results examining the differences between e-scooters operating on roads versus shared segregated cycle lanes. The work concludes with a case study comparing two modelled scenarios, one including a segregated cycle lane and one without. Drawing upon metrics such as the route segmentation/ cut-off rate and average travel comfort, we can more precisely explore how new forms of mobility will influence different kinds of street users in order to better understand the trade-offs associated with different paths forward.

Mobility-as-a-Service: Simulation of Multi-Modal Operations in Low-Density Cities

In the case of the low-density city, empirical evidence continuously demonstrates that transit investment is not a magic bullet. Desirable outcomes are not guaranteed and are often dependent on development density and other urban characteristics. Mobility-as-a-service (MaaS) presents a new approach: a digital platform providing access to multi-modal travel alternatives and totally comprehensive integrated trip-making, planning, and payment services. Review of the literature highlights shortcomings in traditional transportation planning by examining aspects of multi-modal planning such as adoption, parterships, operations, integration, capacity implications, and impact analyses. To enhance the practice of multi-modal planning, the following experiment evaluates various performance measures and inter-modal interactions on International Drive in Orlando, Florida, U.S., via D- and I-optimal experimental designs in a simulated MaaS network. Alternative scenarios are developed comparing varied modal shares across five travel modes: personal vehicles, transit, ridesourcing (or ride-hailing), micro-mobility, and walking. The modal effects are analyzed to highlight the strengths and weakness of each mode under a variety of congestion conditions. While transit enjoys the lowest impact per person, ridesourcing demonstrates adverse effects across all measures. Based on the novel interactions of transit and ridesourcing with directional demand, strategies are outlined for optimizing ridesourcing-transit integration to reduce route travel time, queuing, and overall network delay. The performance impacts of curbside facilities are also discussed for improved multi-modal integration at the street level. These findings are applied to propose a framework for effective planning and implementation of mobility services in low-density cities, focused on operations, city-level connectivity, and curbside management.

Emitting long-distance spiral airborne sound using low-profile planar acoustic antenna

Recent years have witnessed a rapidly growing interest in exploring the use of spiral sound carrying artificial orbital angular momentum (OAM), toward establishing a spiral-wave-based technology that is significantly more efficient in energy or information delivering than the ordinary plane wave technology. A major bottleneck of advancing this technology is the efficient excitation of far-field spiral waves in free space, which is a must in exploring the use of spiral waves for long-distance information transmission and particle manipulation. Here, we report a low-profile planar acoustic antenna to modulate wavefronts emitted from a near-field point source and achieve far-field spiral airborne sound carrying OAM. Using the holographic interferogram as a 2D modulated artificial acoustic impedance metasurface, we show the efficient conversion from the surface wave into the propagating spiral shape beam both numerically and experimentally. The vortex fields with spiral phases originate from the complex inter-modal interactions between cylindrical surface waves and a spatially-modulated impedance boundary condition. This antenna can open new routes to highly integrated spiral sound emitters that are critical for practical acoustic functional devices.

Improved Speaker and Navigator for Vision-and-Language Navigation

Prior works in vision-and-language navigation (VLN) focus on using long short-term memory (LSTM) to carry the flow of information on either the navigation model (navigator) or the instruction generating model (speaker).The outstanding capability of LSTM to process intermodal interactions has been widely verified; however, LSTM neglects the intramodel interactions, leading to negative effect on either navigator or speaker. The performance of attention-based Transformer is satisfactory in sequence-to-sequence translation domains, but Transformer structure implemented directly in VLN has yet been satisfied. In this article, we propose novel Transformer-based multimodal frameworks for the navigator and speaker, respectively. In our frameworks, the multihead self-attention with the residual connection is used to carry the information flow. Specially, we set a switch to prevent them from directly entering the information flow in our navigator framework. In experiments, we verify the effectiveness of our proposed approach, and show significant performance advantages over the baselines.

Strategies for wideband light generation in nonlinear multimode integrated waveguides

In this paper we discuss two strategies to achieve wideband light generation through intermodal nonlinear parametric processes in multimode integrated waveguides. We outline how the interplay among intermodal interactions and high dispersion may lead to the generation of light characterized by substantial power spectral density if compared to supercontinuum sources. These results are valid independently of the waveguide material; however, our numerical simulations focus on silicon waveguides, which are nowadays at the core of integrated photonics. The long-term vision is that by exploiting an adequate number of intermodal processes, widely tunable radiation having high-power spectral density can be generated in a broad portion of the transparency window of silicon.

Multimodal encoders and decoders with gate attention for visual question answering

Visual Question Answering (VQA) is a multimodal research related to Computer Vision (CV) and Natural Language Processing (NLP). How to better obtain useful information from images and questions and give an accurate answer to the question is the core of the VQA task. This paper presents a VQA model based on multimodal encoders and decoders with gate attention (MEDGA). Each encoder and decoder block in the MEDGA applies not only self-attention and crossmodal attention but also gate attention, so that the new model can better focus on inter-modal and intra-modal interactions simultaneously within visual and language modality. Besides, MEDGA further filters out noise information irrelevant to the results via gate attention and finally outputs attention results that are closely related to visual features and language features, which makes the answer prediction result more accurate. Experimental evaluations on the VQA 2.0 dataset and the ablation experiments under different conditions prove the effectiveness of MEDGA. In addition, the MEDGA accuracy on the test-std dataset has reached 70.11%, which exceeds many existing methods.

Multi-Modal Interaction Graph Convolutional Network for Temporal Language Localization in Videos.

This paper focuses on tackling the problem of temporal language localization in videos, which aims to identify the start and end points of a moment described by a natural language sentence in an untrimmed video. However, it is non-trivial since it requires not only the comprehensive understanding of the video and sentence query, but also the accurate semantic correspondence capture between them. Existing efforts are mainly centered on exploring the sequential relation among video clips and query words to reason the video and sentence query, neglecting the other intra-modal relations (e.g., semantic similarity among video clips and syntactic dependency among the query words). Towards this end, in this work, we propose a Multi-modal Interaction Graph Convolutional Network (MIGCN), which jointly explores the complex intra-modal relations and inter-modal interactions residing in the video and sentence query to facilitate the understanding and semantic correspondence capture of the video and sentence query. In addition, we devise an adaptive context-aware localization method, where the context information is taken into the candidate moments and the multi-scale fully connected layers are designed to rank and adjust the boundary of the generated coarse candidate moments with different lengths. Extensive experiments on Charades-STA and ActivityNet datasets demonstrate the promising performance and superior efficiency of our model.

Nonlinear Impairment Scaling in Multi-Mode Fibers for Mode-Division Multiplexing

The scaling of nonlinear effects in multi-mode fibers with the mode count has been investigated by analytical studies and by numerical simulations. Increasing the core radius results in larger effective mode field areas, which reduces the impact of nonlinear effects. On the other hand, the presence of more interaction partners for intermodal nonlinear interactions due to a higher number of propagating modes strengthens the impact of nonlinear signal degradations on data transmission. Our results indicate that these two trends are mostly balanced even for large mode counts. Transmission reaches and system performance comparable to standard single-mode fibers are achievable for at least 100 modes.

Learning Inter-Modal Correspondence and Phenotypes From Multi-Modal Electronic Health Records

Non-negative tensor factorization has been shown a practical solution to automatically discover phenotypes from the electronic health records (EHR) with minimal human supervision. Such methods generally require an input tensor describing the inter-modal interactions to be pre-established; however, the correspondence between different modalities (e.g., correspondence between medications and diagnoses) can often be missing in practice. Although heuristic methods can be applied to estimate them, they inevitably introduce errors, and leads to sub-optimal phenotype quality. This is particularly important for patients with complex health conditions (e.g., in critical care) as multiple diagnoses and medications are simultaneously present in the records. To alleviate this problem and discover phenotypes from EHR with unobserved inter-modal correspondence, we propose the collective hidden interaction tensor factorization (cHITF) to infer the correspondence between multiple modalities jointly with the phenotype discovery. We assume that the observed matrix for each modality is marginalization of the unobserved inter-modal correspondence, which are reconstructed by maximizing the likelihood of the observed matrices. Extensive experiments conducted on the real-world MIMIC-III dataset demonstrate that cHITF effectively infers clinically meaningful inter-modal correspondence, discovers phenotypes that are more clinically relevant and diverse, and achieves better predictive performance compared with a number of state-of-the-art computational phenotyping models.

High-energy multidimensional solitary states in hollow-core fibres

Multidimensional solitary states (MDSS)—self-sustained wavepackets—have attracted renewed interest in many different fields of physics. They are of particular importance in nonlinear optics, especially for the nonlinear propagation of ultrashort pulses in multimode fibres, which contain rich spatiotemporal intermodal interactions and dynamics, albeit often in an unstable manner. Here, we report the observation of the formation of highly stable multidimensional solitary states in a molecular gas-filled large-core hollow-core fibre. We experimentally and numerically demonstrate the creation of MDSS by multimillijoule, subpicosecond near-infrared pulses and the underlying physics. We find that the MDSS have a broadband redshifted spectra with an uncommon negative quadratic spectral phase at the output of the hollow-core fibre, originating from Raman enhancement due to the strong intermodal nonlinear interactions. The spatial and temporal localization of MDSS enables the compression of the broadened pulses at the output to 10.8 fs by simple linear propagation in a piece of fused silica. The high spatiotemporal quality of MDSS is further verified by high-harmonic generation. Our results present new opportunities for studying multimodal spatiotemporal dynamics in the high-energy regime. This work also presents a route toward a new class of compact, tunable and high-energy spatiotemporally engineered coherent light sources based on picosecond ytterbium technology. The formation of multidimensional solitary states through the nonlinear propagation of high-energy pulses in a molecular gas-filled large-core hollow-core fibre is demonstrated, offering new opportunities for studying multimodal spatiotemporal dynamics in the high-energy regime.

OPPORTUNITIES FOR IMPROVING THE COMPETITIVENESS OF RAIL FREIGHT TRANSPORTATION ALONG THE TRANSPORT CORRIDORS OF THE RUSSIAN FEDERATION

The greatest prospects associated with transnational rail freight are determined by the use of containers. In Russian publications, the development of containerized cargo transportation is mainly considered by analyzing the general laws of functioning of international transport corridors. At the same time, aspects affecting infrastructure diagnostics and analysis of the throughput capacity of sections of the railway network are not considered.The purpose of the study was to identify the characteristics of freight flows in the direction of the EU—EAEU—PRC, to identify infrastructure barriers that slow down the acceleration and increase in the volume of container railway freight transportation, as well as to develop a set of measures aimed at improving the competitiveness of container railway freight transportation along the transport corridors of the Russian Federation.In the course of the study, infrastructural diagnostics was carried out, which revealed a number of barriers that did not allow realizing the growth potential of transit container cargo transportation.To remove infrastructure barriers, it is necessary to implement a number of measures, the key of which are measures to increase the capacity of infrastructure at border crossings, increase the speed of trains along the railway networks of JSCo RZD, develop terminal and logistics centers that provide a wide range of services for consignors and consignees.The development of container rail freight is a strategically important area for the Russian Federation. The well-coordinated work of Russian Railways with private investors and the state in the framework of expanding and modernizing the terminal infrastructure, as well as increasing intermodal interactions, is able to overcome the existing infrastructure barriers and ensure the growth of container transit in the required volumes.

Intra- and Inter-modal Multilinear Pooling with Multitask Learning for Video Grounding

Video grounding aims to temporally localize an action in an untrimmed video referred to by a query in natural language, which plays an important role in fine-grained video understanding. Given temporal proposals of limited granularity, the task is challenging that it requires fusing multi-modal features from questions and videos effectively, and localizing the referred action accurately. For multimodal feature fusion, we present an Intra- and Inter-modal Multilinear pooling (IIM) model to effectively combine the multi-modal features with considering both the intra- and inter-modal feature interactions. Compared to existing multimodal fusion models, IIM can capture high-order interactions and is more capable for modeling temporal information of videos. For action localization, we propose a simple yet effective multi-task learning framework to simultaneously predict the action label, alignment score and refined location in an end-to-end manner. Experimental results on real-world TaCoS and Charades-STA datasets demonstrate the superiority of the proposed approach over existing state-of-the-art methods.

Energy Spectrum of Turbulent Velocity Pulsations at Arbitrary Values of Fluid Viscosity

The problem of calculating the energy spectrum of turbulent velocity pulsations in the case of homogeneous isotropic and stationary turbulence is considered. The domain of turbulent energy production is treated as “a black box” on which boundary the spectral energy flux is given. It is assumed that the spectrum is formatted due to intermodal interactions being local in the wave-number space that leads to a cascade mechanism of energy transfer along the wave-number spectrum and the possibility of using the renormalization-group method related to the Markovian features of the process under consideration. The obtained formula for energy spectrum is valid in a wide wave-number range and at arbitrary values of fluid viscosity. It is shown that in functional formulation of the statistical theory of turbulence, the procedure of separating local intermodal interactions, which govern energy transfer (straining effect), and filtering out nonlocal interactions, which have no influence on energy transfer (sweeping effect), is directly described without providing additional arguments or conjectures commonly used in the renormalization-group analysis of turbulent spectra.

Nonlinear interactions in multimode optical fibers

Abstract The limitations of single-mode fibers incite multimode fibers to meet the technological demands and explore new physics. Multimode fibers support many spatial modes which can interact with each other through various nonlinear processes and give rise to complex spatiotemporal dynamics. In this review paper, we address intermodal coupling based nonlinear phenomena in short and long pulse regimes. Ultrashort femtosecond pulse regime is dominated by four-wave mixing, self and cross-phase modulation, interpulse as well as intrapulse Raman scattering and provides strong platform for the realization of multitude of nonlinear phenomena such as multimode soliton formation, generation of intense combs of dispersive waves from the multimode soliton, beam clean-up, geometric parametric instability, and optical event horizon. On the other hand, long pulse regime exhibits strong stimulated Raman processes and provides natural tool for the investigation of nonlinear phenomena such as, cascaded Raman scattering, intermodal four-wave mixing based wideband frequency generation, transient grating formation and many others. This article reviews the recent developments in intermodal nonlinear interactions in multimode fiber for both the short and long pulse regimes and provides a roadmap for future research directions in the area of multimode nonlinear fiber optics.

Multimodal Transformer With Multi-View Visual Representation for Image Captioning

Image captioning aims to automatically generate a natural language description of a given image, and most state-of-the-art models have adopted an encoder-decoder framework. The framework consists of a convolution neural network (CNN)-based image encoder that extracts region-based visual features from the input image, and an recurrent neural network (RNN)-based caption decoder that generates the output caption words based on the visual features with the attention mechanism. Despite the success of existing studies, current methods only model the co-attention that characterizes the inter-modal interactions while neglecting the self-attention that characterizes the intra-modal interactions. Inspired by the success of the Transformer model in machine translation, here we extend it to a Multimodal Transformer (MT) model for image captioning. Compared to existing image captioning approaches, the MT model simultaneously captures intra- and inter-modal interactions in a unified attention block. Due to the in-depth modular composition of such attention blocks, the MT model can perform complex multimodal reasoning and output accurate captions. Moreover, to further improve the image captioning performance, multi-view visual features are seamlessly introduced into the MT model. We quantitatively and qualitatively evaluate our approach using the benchmark MSCOCO image captioning dataset and conduct extensive ablation studies to investigate the reasons behind its effectiveness. The experimental results show that our method significantly outperforms the previous state-of-the-art methods. With an ensemble of seven models, our solution ranks the 1st place on the real-time leaderboard of the MSCOCO image captioning challenge at the time of the writing of this paper.

Reconfigurable generation of optical vortices based on forward stimulated intermodal Brillouin scattering in subwavelength-hole photonic waveguides

An all-optically reconfigurable generation of optical vortices would be highly beneficial to the implementation of next-generation optical communication and advanced information processing. The previously demonstrated approaches based on the parametric nonlinear optical processes, however, have exhibited limited conversion efficiency due to the group velocity mismatch and nonlinear phase shifts, and require the cumbersome preparation of either the optical element or initial seed beam having a non-zero topological charge. Here, we propose and analyze a novel scheme for highly efficient all-optical generation and control of optical vortices based on the dynamic acoustic vortex grating created by forward stimulated intermodal Brillouin scattering in a subwavelength-hole photonic waveguide. The dual-frequency pump beams in two different hybrid optical modes drive an acoustic vortex mode, which transforms a signal in the fundamental optical mode into an optical vortex mode. This scheme not only eliminates the need for the initial preparation of an angular-momentum-carrying medium or an optical vortex seed but also guarantees high modal purity and nearly 100% conversion efficiency assisted by the energy-momentum conservation. We also investigate the feasibility and practicability of the subwavelength-hole waveguides by examining the intermodal conversion efficiency and robustness of guidance of the optical vortices, taking into account the impact of the Kerr-type nonlinear effects on the intermodal Brillouin interactions based on our rigorous full-vectorial analytical theory.

Accelerated nonlinear interactions in graded-index multimode fibers

Multimode optical fibers have recently reemerged as a viable platform for addressing a number of long-standing issues associated with information bandwidth requirements and power-handling capabilities. As shown in recent studies, the complex nature of such heavily multimoded systems can be effectively exploited to observe altogether novel physical effects arising from spatiotemporal and intermodal linear and nonlinear processes. Here, we study for the first time, accelerated nonlinear intermodal interactions in core-diameter decreasing multimode fibers. We demonstrate that in the anomalous dispersion region, this spatiotemporal acceleration can lead to relatively blue-shifted multimode solitons and blue-drifting dispersive wave combs, while in the normal domain, to a notably flat and uniform supercontinuum, extending over 2.5 octaves. Our results pave the way towards a deeper understanding of the physics and complexity of nonlinear, heavily multimoded optical systems, and could lead to highly tunable optical sources with very high spectral densities.