Search

Deep learning is commonly used to reconstruct HDR images from LDR images. LDR stack-based methods are used for single-image HDR reconstruction, generating an HDR image from a deep learning-generated LDR stack. However, current methods generate the stack with predetermined exposure values (EVs), which may limit the quality of HDR reconstruction. To address this, we propose the continuous exposure value representation (CEVR), which uses an implicit function to generate LDR images with arbitrary EVs, including those unseen during training. Our approach generates a continuous stack with more images containing diverse EVs, significantly improving HDR reconstruction. We use a cycle training strategy to supervise the model in generating continuous EV LDR images without corresponding ground truths. Our CEVR model outperforms existing methods, as demonstrated by experimental results., Comment: ICCV 2023. Project page: https://skchen1993.github.io/CEVR_web

In the realm of Tiny AI, we introduce ``You Only Look at Interested Cells" (YOLIC), an efficient method for object localization and classification on edge devices. Through seamlessly blending the strengths of semantic segmentation and object detection, YOLIC offers superior computational efficiency and precision. By adopting Cells of Interest for classification instead of individual pixels, YOLIC encapsulates relevant information, reduces computational load, and enables rough object shape inference. Importantly, the need for bounding box regression is obviated, as YOLIC capitalizes on the predetermined cell configuration that provides information about potential object location, size, and shape. To tackle the issue of single-label classification limitations, a multi-label classification approach is applied to each cell for effectively recognizing overlapping or closely situated objects. This paper presents extensive experiments on multiple datasets to demonstrate that YOLIC achieves detection performance comparable to the state-of-the-art YOLO algorithms while surpassing in speed, exceeding 30fps on a Raspberry Pi 4B CPU. All resources related to this study, including datasets, cell designer, image annotation tool, and source code, have been made publicly available on our project website at https://kai3316.github.io/yolic.github.io

We propose a sparse end-to-end multi-person pose regression framework, termed QueryPose, which can directly predict multi-person keypoint sequences from the input image. The existing end-to-end methods rely on dense representations to preserve the spatial detail and structure for precise keypoint localization. However, the dense paradigm introduces complex and redundant post-processes during inference. In our framework, each human instance is encoded by several learnable spatial-aware part-level queries associated with an instance-level query. First, we propose the Spatial Part Embedding Generation Module (SPEGM) that considers the local spatial attention mechanism to generate several spatial-sensitive part embeddings, which contain spatial details and structural information for enhancing the part-level queries. Second, we introduce the Selective Iteration Module (SIM) to adaptively update the sparse part-level queries via the generated spatial-sensitive part embeddings stage-by-stage. Based on the two proposed modules, the part-level queries are able to fully encode the spatial details and structural information for precise keypoint regression. With the bipartite matching, QueryPose avoids the hand-designed post-processes and surpasses the existing dense end-to-end methods with 73.6 AP on MS COCO mini-val set and 72.7 AP on CrowdPose test set. Code is available at https://github.com/buptxyb666/QueryPose., Comment: Published on NeurIPS 2022

Multi-person pose estimation generally follows top-down and bottom-up paradigms. Both of them use an extra stage ($\boldsymbol{e.g.,}$ human detection in top-down paradigm or grouping process in bottom-up paradigm) to build the relationship between the human instance and corresponding keypoints, thus leading to the high computation cost and redundant two-stage pipeline. To address the above issue, we propose to represent the human parts as adaptive points and introduce a fine-grained body representation method. The novel body representation is able to sufficiently encode the diverse pose information and effectively model the relationship between the human instance and corresponding keypoints in a single-forward pass. With the proposed body representation, we further deliver a compact single-stage multi-person pose regression network, termed as AdaptivePose. During inference, our proposed network only needs a single-step decode operation to form the multi-person pose without complex post-processes and refinements. We employ AdaptivePose for both 2D/3D multi-person pose estimation tasks to verify the effectiveness of AdaptivePose. Without any bells and whistles, we achieve the most competitive performance on MS COCO and CrowdPose in terms of accuracy and speed. Furthermore, the outstanding performance on MuCo-3DHP and MuPoTS-3D further demonstrates the effectiveness and generalizability on 3D scenes. Code is available at https://github.com/buptxyb666/AdaptivePose., Comment: Submit to IEEE TCSVT; 11 pages. arXiv admin note: text overlap with arXiv:2112.13635

Knowledge graph (KG) embeddings have been a mainstream approach for reasoning over incomplete KGs. However, limited by their inherently shallow and static architectures, they can hardly deal with the rising focus on complex logical queries, which comprise logical operators, imputed edges, multiple source entities, and unknown intermediate entities. In this work, we present the Knowledge Graph Transformer (kgTransformer) with masked pre-training and fine-tuning strategies. We design a KG triple transformation method to enable Transformer to handle KGs, which is further strengthened by the Mixture-of-Experts (MoE) sparse activation. We then formulate the complex logical queries as masked prediction and introduce a two-stage masked pre-training strategy to improve transferability and generalizability. Extensive experiments on two benchmarks demonstrate that kgTransformer can consistently outperform both KG embedding-based baselines and advanced encoders on nine in-domain and out-of-domain reasoning tasks. Additionally, kgTransformer can reason with explainability via providing the full reasoning paths to interpret given answers., Comment: kgTransformer; Accepted to KDD 2022

Supervised learning is dominant in person search, but it requires elaborate labeling of bounding boxes and identities. Large-scale labeled training data is often difficult to collect, especially for person identities. A natural question is whether a good person search model can be trained without the need of identity supervision. In this paper, we present a weakly supervised setting where only bounding box annotations are available. Based on this new setting, we provide an effective baseline model termed Region Siamese Networks (R-SiamNets). Towards learning useful representations for recognition in the absence of identity labels, we supervise the R-SiamNet with instance-level consistency loss and cluster-level contrastive loss. For instance-level consistency learning, the R-SiamNet is constrained to extract consistent features from each person region with or without out-of-region context. For cluster-level contrastive learning, we enforce the aggregation of closest instances and the separation of dissimilar ones in feature space. Extensive experiments validate the utility of our weakly supervised method. Our model achieves the rank-1 of 87.1% and mAP of 86.0% on CUHK-SYSU benchmark, which surpasses several fully supervised methods, such as OIM and MGTS, by a clear margin. More promising performance can be reached by incorporating extra training data. We hope this work could encourage the future research in this field., Comment: Accepted by ICCV 2021

Video scene parsing is a long-standing challenging task in computer vision, aiming to assign pre-defined semantic labels to pixels of all frames in a given video. Compared with image semantic segmentation, this task pays more attention on studying how to adopt the temporal information to obtain higher predictive accuracy. In this report, we introduce our solution for the 1st Video Scene Parsing in the Wild Challenge, which achieves a mIoU of 57.44 and obtained the 2nd place (our team name is CharlesBLWX)., Comment: technical report for "The 1st Video Scene Parsing in the Wild Challenge Workshop". arXiv admin note: text overlap with arXiv:2108.11819

Multi-Person Pose Estimation is an interesting yet challenging task in computer vision. In this paper, we conduct a series of refinements with the MSPN and PoseFix Networks, and empirically evaluate their impact on the final model performance through ablation studies. By taking all the refinements, we achieve 78.7 on the COCO test-dev dataset and 76.3 on the COCO test-challenge dataset.

Semi-supervised video object segmentation is an interesting yet challenging task in machine learning. In this work, we conduct a series of refinements with the propagation-based video object segmentation method and empirically evaluate their impact on the final model performance through ablation study. By taking all the refinements, we improve the space-time memory networks to achieve a Overall of 79.1 on the Youtube-VOS Challenge 2019.

Multi-person pose estimation is a fundamental yet challenging task in computer vision. Both rich context information and spatial information are required to precisely locate the keypoints for all persons in an image. In this paper, a novel Context-and-Spatial Aware Network (CSANet), which integrates both a Context Aware Path and Spatial Aware Path, is proposed to obtain effective features involving both context information and spatial information. Specifically, we design a Context Aware Path with structure supervision strategy and spatial pyramid pooling strategy to enhance the context information. Meanwhile, a Spatial Aware Path is proposed to preserve the spatial information, which also shortens the information propagation path from low-level features to high-level features. On top of these two paths, we employ a Heavy Head Path to further combine and enhance the features effectively. Experimentally, our proposed network outperforms state-of-the-art methods on the COCO keypoint benchmark, which verifies the effectiveness of our method and further corroborates the above proposition.

Multi-person pose estimation is an important but challenging problem in computer vision. Although current approaches have achieved significant progress by fusing the multi-scale feature maps, they pay little attention to enhancing the channel-wise and spatial information of the feature maps. In this paper, we propose two novel modules to perform the enhancement of the information for the multi-person pose estimation. First, a Channel Shuffle Module (CSM) is proposed to adopt the channel shuffle operation on the feature maps with different levels, promoting cross-channel information communication among the pyramid feature maps. Second, a Spatial, Channel-wise Attention Residual Bottleneck (SCARB) is designed to boost the original residual unit with attention mechanism, adaptively highlighting the information of the feature maps both in the spatial and channel-wise context. The effectiveness of our proposed modules is evaluated on the COCO keypoint benchmark, and experimental results show that our approach achieves the state-of-the-art results., Comment: Accepted by CVPR 2019

The mechanical properties of the microstructures surrounding cells influence the behavior of cells in differentiation, proliferation, and apoptosis, etc. The stiffness of the extra-cellular microenvironment has been shown to be one such mechanical property. Studies reported in the literature concerning the stiffness of the extracellular microenvironment mainly sought to understand the scientific principles and mechanisms underlying its effect on cell-environment interaction. This paper describes an approach that achieves such manipulation, and reports experimental results that demonstrate the effectiveness of this proposed approach.

In this paper, an active vibration isolation system based on hybrid algorithm is presented in a wide frequency band. Initially, a nonlinear magnetostrictive actuator model is used to establish the appropriate parameters by experiments, which make the actuator using in vibration isolation system work in a better linear dynamic performance, then the sliding mode algorithm modified by neural network, a hybrid algorithm is proposed as the active control controller, and its stability is also analyzed by Lyapunov theory. Furthermore, a dynamic virtual prototype model of active vibration isolation is established to carry out the co-simulation with Adams and Matlab/Simulink, and the results show that under the difference excitations, the neural network sliding mode controller exhibits a good control performance, and the active vibration isolation can effectively improve the vibration isolation effect, reduce the force transmitted to the base and broaden the vibration isolation bandwidth.

Eolian dust plays an important role in the Earth’s climate system. Environmental magnetism has been widely used to trace dust variations at different spatial and temporal scales. However, the magnetic properties of sediments from key dust sources have not been well determined. In this study, surface samples from potential dust sources in inner Eastern Asia were systematically investigated. Our results indicate that ferrimagnetic and antiferromagnetic minerals are both present in surface sediments and that they have broad grain size distributions. Ferrimagnetic components are dominated by partially oxidized coarse-grained (pseudo-single domain and multi-domain) lithogenic magnetite particles with minor contributions from pedogenic fine-grained (single domain and superparamagnetic) particles. Antiferromagnetic hematite can be classified into three groups in terms of diffuse reflectance spectroscopy (DRS) band positions (P 560 nm, P 545 nm, and P 535 nm, where numbers indicate the DRS band wavelength for hematite). The first group (P 560 nm) is the coarse-grained hematite of lithogenic origin and is mostly confined to western China. The P 535 nm group is of pedogenic origin. The P 545 nm group is an intermediate phase that is present both in surface samples from the source regions and in loess. Therefore, the P 545 nm and P 535 nm groups are related to eolian inputs to the Chinese Loess Plateau and pedogenic processes, respectively. In addition, significant differences exist between the magnetic properties of eolian material from sources and depositional regions due to gravitational sorting. These insights provide strong constraints on interpretation of dust signals recorded by the Chinese loess and marine sediments from the North Pacific Ocean.

With the increasing interest in connected vehicles, it is useful to evaluate the capability of delivering large content over a WiFi infrastructure to vehicles. The throughput achieved over WiFi channels can be highly variable and also rapidly degrades as the distance from the access point increases. While this behavior is well understood at the data link layer, the interactions across the various protocol layers (data link and up through the transport layer) and the effect of mobility may reduce the amount of content transferred to the vehicle, as it travels along the roadway. This paper examines the throughput achieved at the TCP layer over a carefully designed outdoor WiFi environment and the interactions across the layers that impact the performance achieved, as a function of the receiver mobility. The experimental studies conducted reveal that impairments over the WiFi link (frame loss, ARQ and increased delay) and the residual loss seen by TCP causes a cascade of duplicate ACKs to be generated. This triggers large congestion window reductions at the sender, leading to a drastic degradation of throughput to the vehicular client. To ensure outdoor WiFi infrastructures have the potential to sustain reasonable downlink throughput for drive-by vehicles, we speculate that there is a need to adapt how WiFi and TCP (as well as mobility protocols) function for such vehicular applications.

Current Internet performs traffic engineering (TE) by estimating traffic matrices on a regular schedule, and allocating flows based upon weights computed from these matrices. This means the allocation is based upon a guess of the traffic in the network based on its history. Information-Centric Networks on the other hand provide a finer-grained description of the traffic: a content between a client and a server is uniquely identified by its name, and the network can therefore learn the size of different content items, and perform traffic engineering and resource allocation accordingly. We claim that Information-Centric Networks can therefore provide a better handle to perform traffic engineering, resulting in significant performance gain. We present a mechanism to perform such resource allocation. We see that our traffic engineering method only requires knowledge of the flow size (which, in ICN, can be learned from previous data transfers) and outperforms a min-MLU allocation in terms of response time. We also see that our method identifies the traffic allocation patterns similar to that of min-MLU without having access to the traffic matrix ahead of time. We show a very significant gain in response time where min MLU is almost 50% slower than our ICN-based TE method.

Two main trends in today's internet are of major interest for video streaming services: most content delivery platforms coincide towards using adaptive video streaming over HTTP and new network architectures allowing caching at intermediate points within the network. We investigate one of the most popular streaming service in terms of rate adaptation and opportunistic caching. Our experimental study shows that the streaming client's rate selection trajectory, i.e., the set of selected segments of varied bit rates which constitute a complete video, is not repetitive across separate downloads. Also, the involvement of caching could lead to frequent alternation between cache and server when serving back client's requests for video segments. These observations warrant cautions for rate adaption algorithm design and trigger our analysis to characterize the performance of in-network caching for HTTP streaming. Our analytic results show: (i) a significant degradation of cache hit rate for adaptive streaming, with a typical file popularity distribution in nowadays internet; (ii) as a result of the (usually) higher throughput at the client-cache connection compared to client-server one, cache-server oscillations due to misjudgments of the rate adaptation algorithm occur. Finally, we introduce DASH-INC, a framework for improved video streaming in caching networks including transcoding and multiple throughput estimation.

Resource allocation in wireless networks typically occurs at PHY/MAC layers, while random network coding (RNC) is a network layer strategy. An interesting question is how resource allocation mechanisms can be tuned to improve RNC performance. By means of a differential equation framework which models RNC throughput in terms of lower layer parameters, we propose a gradient based approach that can dynamically allocate MAC and PHY layer resources with the goal of maximizing the minimum network coding throughput among all the destination nodes in a RNC multicast. We exemplify this general approach with two resource allocation problems: (i) power control to improve network coding throughput, and (ii) CSMA mean backoff delay control to improve network coding throughput. We design both centralized algorithms and online algorithms for power control and CSMA backoff control. Our evaluations, including numerically solving the differential equations in the centralized algorithm and an event-driven simulation for the online algorithm, show that such gradient based dynamic resource allocation yields significant throughput improvement of the destination nodes in RNC. Further, our numerical results reveal that network coding aware power control can regain the broadcast advantage of wireless transmissions to improve the throughput., Comment: 28 pages, 9 figures. Submitted to IEEE Transactions on Wireless Communications

Protease activity measurement has broad application in drug screening, diagnosis and disease staging, and molecular profiling. However, conventional immunopeptidemetric assays (IMPA) exhibit low fluorescence signal-to-noise ratios, preventing reliable measurements at lower concentrations in the clinically important picomolar to nanomolar range. Here, we demonstrated a highly sensitive measurement of protease activity using a nanoplasmonic resonator (NPR). NPRs enhance Raman signals by 6.1 x 10(exp 10) times in a highly reproducible manner, enabling fast detection of proteolytically active prostate-specific antigen (paPSA) activities in real-time, at a sensitivity level of 6 pM (0.2 ng/mL) with a dynamic range of 3 orders of magnitude. Experiments on extracellular fluid (ECF) from the paPSA-positive cells demonstrate specific detection in a complex biofluid background. This method offers a fast, sensitive, accurate, and one-step approach to detect the proteases' activities in very small sample volumes., Accepted 25 Jan 2010 for publication in ACS Nano; final version published in v4 n2 p978-984, 23 Feb 2010.

Thesis (S.M.)--Massachusetts Institute of Technology, Engineering Systems Division, Technology and Policy Program; and, (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2004., Includes bibliographical references (p. 64-69)., Databases and analysis tools currently being used to study carbon dioxide capture and storage (CCS) options are managed by diverse organizations and are heterogeneous in format. Tools to study the various components of a CCS system have been developed in several fields including chemistry, geology, and economics. Data being used to run analyses are being obtained from an equally diverse set of organizations, from data collected for environmental assessments to data on oil and gas exploration. These variations in tools and data cause complications in systems-level analyses, resulting in additional effort expended in data collection and opportunities for human error. A geographic information system has been implemented to automate and support robust studies of both component and system options. Context management and information integration techniques have been designed into the system. The system improves the availability and quality of information by automatically managing the distributed and heterogeneous data sources. The resulting information is being used to advance research and development of CCS systems through efforts such as the NETL sponsored Regional Carbon Sequestration Partnerships. This paper will present an overview of the system and initial results of its application to CCS-related data., by David Su-Kai Cheng., S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, Engineering Systems Division, Technology and Policy Program; and, (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2004., Includes bibliographical references (p. 64-69)., Databases and analysis tools currently being used to study carbon dioxide capture and storage (CCS) options are managed by diverse organizations and are heterogeneous in format. Tools to study the various components of a CCS system have been developed in several fields including chemistry, geology, and economics. Data being used to run analyses are being obtained from an equally diverse set of organizations, from data collected for environmental assessments to data on oil and gas exploration. These variations in tools and data cause complications in systems-level analyses, resulting in additional effort expended in data collection and opportunities for human error. A geographic information system has been implemented to automate and support robust studies of both component and system options. Context management and information integration techniques have been designed into the system. The system improves the availability and quality of information by automatically managing the distributed and heterogeneous data sources. The resulting information is being used to advance research and development of CCS systems through efforts such as the NETL sponsored Regional Carbon Sequestration Partnerships. This paper will present an overview of the system and initial results of its application to CCS-related data., by David Su-Kai Cheng., S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, Engineering Systems Division, Technology and Policy Program; and, (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2004., Includes bibliographical references (p. 64-69)., Databases and analysis tools currently being used to study carbon dioxide capture and storage (CCS) options are managed by diverse organizations and are heterogeneous in format. Tools to study the various components of a CCS system have been developed in several fields including chemistry, geology, and economics. Data being used to run analyses are being obtained from an equally diverse set of organizations, from data collected for environmental assessments to data on oil and gas exploration. These variations in tools and data cause complications in systems-level analyses, resulting in additional effort expended in data collection and opportunities for human error. A geographic information system has been implemented to automate and support robust studies of both component and system options. Context management and information integration techniques have been designed into the system. The system improves the availability and quality of information by automatically managing the distributed and heterogeneous data sources. The resulting information is being used to advance research and development of CCS systems through efforts such as the NETL sponsored Regional Carbon Sequestration Partnerships. This paper will present an overview of the system and initial results of its application to CCS-related data., by David Su-Kai Cheng., S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, Engineering Systems Division, Technology and Policy Program; and, (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2004., Includes bibliographical references (p. 64-69)., Databases and analysis tools currently being used to study carbon dioxide capture and storage (CCS) options are managed by diverse organizations and are heterogeneous in format. Tools to study the various components of a CCS system have been developed in several fields including chemistry, geology, and economics. Data being used to run analyses are being obtained from an equally diverse set of organizations, from data collected for environmental assessments to data on oil and gas exploration. These variations in tools and data cause complications in systems-level analyses, resulting in additional effort expended in data collection and opportunities for human error. A geographic information system has been implemented to automate and support robust studies of both component and system options. Context management and information integration techniques have been designed into the system. The system improves the availability and quality of information by automatically managing the distributed and heterogeneous data sources. The resulting information is being used to advance research and development of CCS systems through efforts such as the NETL sponsored Regional Carbon Sequestration Partnerships. This paper will present an overview of the system and initial results of its application to CCS-related data., by David Su-Kai Cheng., S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, Engineering Systems Division, Technology and Policy Program; and, (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2004., Includes bibliographical references (p. 64-69)., Databases and analysis tools currently being used to study carbon dioxide capture and storage (CCS) options are managed by diverse organizations and are heterogeneous in format. Tools to study the various components of a CCS system have been developed in several fields including chemistry, geology, and economics. Data being used to run analyses are being obtained from an equally diverse set of organizations, from data collected for environmental assessments to data on oil and gas exploration. These variations in tools and data cause complications in systems-level analyses, resulting in additional effort expended in data collection and opportunities for human error. A geographic information system has been implemented to automate and support robust studies of both component and system options. Context management and information integration techniques have been designed into the system. The system improves the availability and quality of information by automatically managing the distributed and heterogeneous data sources. The resulting information is being used to advance research and development of CCS systems through efforts such as the NETL sponsored Regional Carbon Sequestration Partnerships. This paper will present an overview of the system and initial results of its application to CCS-related data., by David Su-Kai Cheng., S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, Engineering Systems Division, Technology and Policy Program; and, (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2004., Includes bibliographical references (p. 64-69)., Databases and analysis tools currently being used to study carbon dioxide capture and storage (CCS) options are managed by diverse organizations and are heterogeneous in format. Tools to study the various components of a CCS system have been developed in several fields including chemistry, geology, and economics. Data being used to run analyses are being obtained from an equally diverse set of organizations, from data collected for environmental assessments to data on oil and gas exploration. These variations in tools and data cause complications in systems-level analyses, resulting in additional effort expended in data collection and opportunities for human error. A geographic information system has been implemented to automate and support robust studies of both component and system options. Context management and information integration techniques have been designed into the system. The system improves the availability and quality of information by automatically managing the distributed and heterogeneous data sources. The resulting information is being used to advance research and development of CCS systems through efforts such as the NETL sponsored Regional Carbon Sequestration Partnerships. This paper will present an overview of the system and initial results of its application to CCS-related data., by David Su-Kai Cheng., S.M.

Thesis (S.M.)--Massachusetts Institute of Technology, Engineering Systems Division, Technology and Policy Program; and, (S.M.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2004., Includes bibliographical references (p. 64-69)., Databases and analysis tools currently being used to study carbon dioxide capture and storage (CCS) options are managed by diverse organizations and are heterogeneous in format. Tools to study the various components of a CCS system have been developed in several fields including chemistry, geology, and economics. Data being used to run analyses are being obtained from an equally diverse set of organizations, from data collected for environmental assessments to data on oil and gas exploration. These variations in tools and data cause complications in systems-level analyses, resulting in additional effort expended in data collection and opportunities for human error. A geographic information system has been implemented to automate and support robust studies of both component and system options. Context management and information integration techniques have been designed into the system. The system improves the availability and quality of information by automatically managing the distributed and heterogeneous data sources. The resulting information is being used to advance research and development of CCS systems through efforts such as the NETL sponsored Regional Carbon Sequestration Partnerships. This paper will present an overview of the system and initial results of its application to CCS-related data., by David Su-Kai Cheng., S.M.

The major goal of the project was to develop an adaptive optical system featuring a radical, dynamic and reversible change in focal length using polydimethylsiloxane (silicone) science and technology. The assertion that silicone-functional mesogens (liquid crystals), where the mesogen displayed an intrinsically high birefringence, would behave similarly and yield the necessary nematic liquid crystalline mesophase in an appropriate temperature range, was demonstrated to be a robust hypothesis. The silicone portion of the materials allowed for the tailoring of both the phase behavior and miscibility to comply to the demands of the adaptive optics application. Classes of silicone mesogens, where the mesogen was either a metallomesogen or else a highly pi-conjugated mesogen were successfully prepared. The metallomesogens did not display the required high birefringence (0.2 < Delta n < 0.5); however, many of the conjugated mesogens did. We conclude that for the metallomesogens, the birefringence of the materials is an extremely sensitive function of the molecular structure of the material. Although many of the materials investigated demonstrated smectic phases, we were able to prepare some showing the required nematic phase. Dispersions of select mesogens in a curable silicone polymer were prepared and demonstrated an elastomeric rheology allowing for the fabrication of mechanically- and electrically-active lenses. A crude adaptive optical device was prepared using these lenses which demonstrated a dynamic, scalable and reversible focal length., Prepared in cooperation with Dow Corning Ltd., South Wales, U.K., Dow Corning Corp., Midland, MI and Cambridge University, Dept. of Engineering, Cambridge, U.K.

Boron-containing molecular clusters, polymers and solid state materials show a number of important bonding, structural, chemical and physical properties. However, one of the major limitations to the development and applications of these materials has been the lack of general synthetic methods by which a specific compound or material can be achieved in high yield or in a desired form. Thus, the overall goal of this thesis work was to develop new efficient synthetic routes to these boron-containing compounds. Part I. New synthetic routes to boron-containing polymers and solid state materials. Processable poly(B-vinylborazine) and random poly(styrene-co-B-vinylborazine) polymers were prepared by solution free radical polymerization techniques. Molecular weight studies of these polymers suggested polymer structures consisting of linear and branched polymer chains. The synthesis of new borazine-substituted hydridopolysilazane polymers was achieved by the thermal reaction of borazine with hydridopolysilazane. The modified polymers were converted in high ceramic and chemical yields to SiNCB ceramic composites that exihibit enhanced thermal stabilities and reduced crystallinities. A new method for the syntheses of metal borides and metal-boride/metal-nitride composites that employs processable metal-oxide or metal powder dispersions in a suitable boron-based polymer was developed. This method has been successfully used to generate metal-boride ceramics in various forms, including powders, compacts and films. Part II. New synthetic routes to molecular borane clusters. A new synthetic strategy for achieving monocarbon or dicarbon insertions into polyhedral boron cage compounds, involving the reactions of polarized terminal alkynes with polyhedral borane anions, was developed. The use of this approach allowed the synthesis of a number of new types of carborane clusters including di-, tri- and tetracarbon carboranes. The syntheses of the first seven-vertex nido-carborane zwitte

The authors have reviewed the basic theories of superconductivity and have evaluated Little's proposed organic superconducting polyene from a synthetic and technical point of view. In addition, they have discussed the conductivity of charge transfer type polymers and conformational effects on conductivity. Various biopolymers and vinyl polymers were synthesized with side chains capable of forming charge transfer complexes. A 1-nylon type intramolecule was developed as a charge transfer copolymer. (Author)