We develop a novel approach to explain why AdaBoost is a successful classifier. By introducing a measure of the influence of the noise points (ION) in the training data for the binary classification problem, we prove that there is a strong connection between the IONand the test error.We further identify that the ION of AdaBoost decreases as the iteration number or the complexity of the base learners increases.We confirm that it is impossible to obtain a consistent classifier without deep trees as the base learners of AdaBoost in some complicated situations.We apply AdaBoost in portfolio management via empirical studies in the Chinese market, which corroborates our theoretical propositions.
Recently, a novel retina-inspired camera, namely spike camera, has shown great potential for recording high-speed dynamic scenes. Unlike the conventional digital cameras that compact the visual information within the exposure interval into a single snapshot, the spike camera continuously outputs binary spike streams to record the dynamic scenes, yielding a very high temporal resolution. Most of the existing reconstruction methods for spike camera focus on reconstructing images with the same resolution as spike camera. However, as a trade-off of high temporal resolution, the spatial resolution of spike camera is limited, resulting in inferior details of the reconstruction. To address this issue, we develop a spike camera super-resolution framework, aiming to super resolve high-resolution intensity images from the low-resolution binary spike streams. Due to the relative motion between the camera and the objects to capture, the spikes fired by the same sensor pixel no longer describes the same points in the external scene. In this paper, we properly exploit the relative motion and derive the relationship between light intensity and each spike, so as to recover the external scene with both high temporal and high spatial resolution. Experimental results demonstrate that the proposed method can reconstruct pleasant high-resolution images from low-resolution spike streams.
A novel solar active AgBr/BiOBr/TiO2 catalyst was synthesized by a facile coprecipitation method for solar-driven water remediation. The synthesized material composed of flower-like TiO2 nanoparticles loaded on BiOBr nanosheets and with homogeneous surface distributed Ag/AgBr nanoparticles. The internal electric field between BiOBr/TiO2 heterojunction greatly facilitated the charge carrier migration; the introduction of narrow band gap semiconductors (AgBr and BiOBr) promoted the visible light adsorption; and the Ag/AgBr nanoparticles acted as photosensitizer to further improve the light utilization. The new material showed 7.6- and 4.0-times activity of pure TiO2 and BiOBr under solar light, and the contribution of reactive species on anthracene degradation followed the order of h+ >O2•−> •OH. The degradation mechanism and pathway were proposed based on intermediates analysis and DFT calculation. The QSAR analysis revealed that the environmental risks of contaminants were greatly reduced during the photocatalysis process but some intermediates were still toxic. The high photocatalytic activity, stability and adaptability all indicated that this new material owns great application potential for cost-effective photocatalytic remediation of persistent organic contaminants under solar light.
Abstract Intense and frequent new particle formation (NPF) events have been observed in polluted urban environments, yet the dominant mechanisms are still under debate. To understand the key species and governing processes of NPF in polluted urban environments, we conducted comprehensive measurements in downtown Beijing during January – March 2018. We performed detailed analyses on sulfuric acid cluster composition and budget, as well as the chemical and physical properties of oxidized organic molecules. Our results demonstrate that the fast clustering of sulfuric acid (H2SO4) and base molecules triggered the NPF events, and oxidized organic molecules (OOMs) further helped grow the newly formed particles towards climate- and health-relevant sizes. This synergistic role of H2SO4, base species, and OOMs in NPF is likely representative of polluted urban environments where abundant H2SO4 and base species usually co-exist, and OOMs are with moderately low volatility when produced under high NOx concentrations.
Stability of carbon dioxide foam attracts huge attentions while being of great challenge to many industrial practices. In this paper, aqueous CO2 foam stabilized with synergy of hydrophilic nanoparticles and nonionic surfactants was experimentally investigated at elevated temperatures and extreme salinities. The foam formula consisting of 1.5 wt% T40 and 2.49 mM C12E23 was determined under elevated temperature and salinity conditions. At 80 °C, the foam volume of the C12E23/T40 foam is 225 mL, and the half-life is 32 min, which is 13.9 times that of the C12E23 foam. At a salinity of 17 × 104 mg/L, the foam volume is 185 mL, and the half-life is 71 min, which is 24.4 times that of the C12E23 foam. With increasing salinity and temperature, the interfacial tension rises, and the viscoelastic modulus gradually declines, resulting in a lower foam stability. However, the T40/C12E23 foam has a good temperature and salinity tolerance, which can be used under harsh reservoir conditions. In a heterogeneous microscopic visualization model, water channeling in high-permeability regions inhibits the further increase in oil recovery. Subsequent injection of the C12E23/T40 foam improves the oil recovery factor up to 86.2%. The C12E23/T40 foam enhances the oil recovery by increasing the sweep area and flooding efficiency. In a sandpack model, the plugging pressure gradient of the CO2 foam stabilized by 2.49 mM C12E23 and 1.5 wt% T40 reaches 25.3 MPa/m, which is 12.65 times higher than that of the C12E23 foam. The composite reinforced foam attains good water blocking and profile control effects, thereby increasing the oil recovery factor 20.1% after water flooding. This study of great importance to improve not only the oil recovery efficiency, also anywhere where CO2-in brine foam applicable.
In passive millimeter wave radiation images, reflection will affect the analysis of the target and scene. Based on the theoretical model of reflection, this paper analyzes the change of brightness temperature(TB) in the target region and the reflected region under specific scenes. A target detection method based on the change of TB is proposed to distinguish the target and its reflection region and obtain the target outline. The effectiveness of the method is verified by simulation and imaging experiments.
In-Memory Computing (IMC), which takes advantage of analog multiplication-accumulation (MAC) insides memory, is promising to alleviate the Von-Neumann bottleneck and improve the energy efficiency of deep neural networks (DNNs). Since the time-domain (TD) computing is also an energy-efficient analog computing paradigm, we present an 8kb mixed-signal IMC macro, TD-SRAM, by combining IMC with TD computing. A dual-edge single input (DESI) TD computing topology is proposed, which can significantly improve the area and power efficiencies of TD cell. The TD-SRAM bitcell consisting of a 6T DESI based TD cell and a 6T-SRAM cell supports binary DNNs. In the IMC mode, 60 columns work in parallel and 96-input binary-MAC operations are processed in each column. Implemented in a standard 40-nm CMOS process, the TD-SRAM achieves the high energy efficiency of 537 TOPS/W at 0.9-V supply. With different DNN topologies, the test chips achieve the accuracy of 95.90%-98.00% with a dual 2-bit time-to-digital converter (TDC) in the MNIST dataset.
Passive radiative cooling technology uses the infrared atmospheric window to allow outer space to be a cold sink for heat. However, this effect is one that is only helpful for energy savings in the warmer months. Wang et al. and Tang et al. used the metal-insulator transition in tungsten-doped vanadium dioxide to create window glass and a rooftop coating that circumvents this problem by turning off the radiative cooling at lower temperatures. Because the transition is simply temperature dependent, this effect also happens passively. Model simulations suggest that these materials would lead to energy savings year-round across most of the climate zones in the United States. —BG A smart radiative coating automatically switches thermal radiation power in response to ambient temperature. The sky is a natural heat sink that has been extensively used for passive radiative cooling of households. A lot of focus has been on maximizing the radiative cooling power of roof coating in the hot daytime using static, cooling-optimized material properties. However, the resultant overcooling in cold night or winter times exacerbates the heating cost, especially in climates where heating dominates energy consumption. We approached thermal regulation from an all-season perspective by developing a mechanically flexible coating that adapts its thermal emittance to different ambient temperatures. The fabricated temperature-adaptive radiative coating (TARC) optimally absorbs the solar energy and automatically switches thermal emittance from 0.20 for ambient temperatures lower than 15°C to 0.90 for temperatures above 30°C, driven by a photonically amplified metal-insulator transition. Simulations show that this system outperforms existing roof coatings for energy saving in most climates, especially those with substantial seasonal variations.