Margin theory provides one of the most popular explanations to the success of AdaBoost, where the central point lies in the recognition that margin is the key for characterizing the performance of ...AdaBoost. This theory has been very influential, e.g., it has been used to argue that AdaBoost usually does not overfit since it tends to enlarge the margin even after the training error reaches zero. Previously the minimum margin bound was established for AdaBoost, however, Breiman (1999) 9 pointed out that maximizing the minimum margin does not necessarily lead to a better generalization. Later, Reyzin and Schapire (2006) 37 emphasized that the margin distribution rather than minimum margin is crucial to the performance of AdaBoost. In this paper, we first present the kth margin bound and further study on its relationship to previous work such as the minimum margin bound and Emargin bound. Then, we improve the previous empirical Bernstein bounds (Audibert et al. 2009; Maurer and Pontil, 2009) 2,30, and based on such findings, we defend the margin-based explanation against Breimanʼs doubts by proving a new generalization error bound that considers exactly the same factors as Schapire et al. (1998) 39 but is sharper than Breimanʼs (1999) 9 minimum margin bound. By incorporating factors such as average margin and variance, we present a generalization error bound that is heavily related to the whole margin distribution. We also provide margin distribution bounds for generalization error of voting classifiers in finite VC-dimension space.
Population growth and industrial development have exacerbated environmental pollution of both land and aquatic environments with toxic and harmful materials. Luminescence-based chemical sensors ...crafted for specific hazardous substances operate on host-guest interactions, leading to the detection of target molecules down to the nanomolar range. Particularly, the luminescence-based sensors constructed on the basis of metal-organic frameworks (MOFs) are of increasing interest, as they can not only compensate for the shortcomings of traditional detection techniques, but also can provide more sensitive detection for analytes. Recent years have seen MOFs-based fluorescent sensors show outstanding advantages in the field of hazardous substance identification and detection. Here, we critically discuss the application of MOFs for the detection of a broad scope of hazardous substances, including hazardous gases, heavy metal ions, radioactive ions, antibiotics, pesticides, nitro-explosives, and some harmful solvents as well as luminous and sensing mechanisms of MOF-based fluorescent sensors. The outlook and several crucial issues of this area are also discussed, with the expectation that it may help arouse widespread attention on exploring fluorescent MOFs (LMOFs) in potential sensing applications.
A dipolarizing flux bundle (DFB) is a small magnetotail flux tube (typically < ~3 RE in XGSM and YGSM) with a significantly more dipolar magnetic field than its background. Dipolarizing flux bundles ...typically propagate earthward at a high speed from the near‐Earth reconnection region. Knowledge of a DFB's flux transport properties leads to better understanding of near‐Earth (X = −6 to −30 RE) magnetotail flux transport and thus conversion of magnetic energy to kinetic and thermal plasma energy following magnetic reconnection. We explore DFB properties with a statistical study using data from the Time History of Events and Macroscale Interactions during Substorms mission. To establish the importance of DFB flux transport, we compare it with transport by bursty bulk flows (BBFs) that typically envelop DFBs. Because DFBs coexist with flow bursts inside BBFs, they contribute >65% of BBF flux transport, even though they last only ~30% as long as BBFs. The rate of DFB flux transport increases with proximity to Earth and to the premidnight sector, as well as with geomagnetic activity and distance from the neutral sheet. Under the latter two conditions, the total flux transport by a typical DFB also increases. Dipolarizing flux bundles appear more often during increased geomagnetic activity. Since BBFs have been previously shown to be the major flux transporters in the tail, we conclude that DFBs are the dominant drivers of this transport. The occurrence rate of DFBs as a function of location and geomagnetic activity informs us about processes that shape global convection and energy conversion.
Key Points
Dipolarizing flux bundles are the major flux carrier of bursty bulk flows
DFBs transport flux faster closer to Earth and in tail's premidnight sector
DFBs transport more flux during higher substorm activity
Fe/N/C is a promising non‐Pt electrocatalyst for the oxygen reduction reaction (ORR), but its catalytic activity is considerably inferior to that of Pt in acidic medium, the environment of polymer ...electrolyte membrane fuel cells (PEMFCs). An improved Fe/N/C catalyst (denoted as Fe/N/C‐SCN) derived from Fe(SCN)3, poly‐m‐phenylenediamine, and carbon black is presented. The advantage of using Fe(SCN)3 as iron source is that the obtained catalyst has a high level of S doping and high surface area, and thus exhibits excellent ORR activity (23 A g−1 at 0.80 V) in 0.1 M H2SO4 solution. When the Fe/N/C‐SCN was applied in a PEMFC as cathode catalyst, the maximal power density could exceed 1 W cm−2.
A non‐precious Fe/N/C electrocatalyst was prepared through pyrolysis of Fe(SCN)3, poly‐m‐phenylenediamine, and carbon black. The obtained Fe/N/C catalyst has high level of S doping and high surface area, and thus exhibits excellent catalytic activity for the oxygen reduction reaction in acidic solution. A polymer electrolyte membrane fuel cell using this catalyst as the cathode can yield a maximal power density as high as 1.03 W cm−2.
Electricity markets must match real-time supply and demand of electricity. With increasing penetration of renewable resources, it is important that this balancing is done effectively, considering the ...high uncertainty of wind and solar energy. Storing electrical energy can make the grid more reliable and efficient and energy storage is proposed as a complement to highly variable renewable energy sources. However, for investments in energy storage to increase, participating in the market must become economically viable for owners. This paper proposes a stochastic formulation of a storage owner's arbitrage profit maximization problem under uncertainty in day-ahead and real-time market prices. The proposed model helps storage owners in market bidding and operational decisions and in estimation of the economic viability of energy storage. Case study results on realistic market price data show that the novel stochastic bidding approach does significantly better than the deterministic benchmark.
The gastrointestinal (GI) microbiota of vertebrates plays critical roles in nutrition, development, immunity and resistance against invasive pathogens. In the past decade, research of the GI ...microbiota of mammals has drastically increased our knowledge on the microbiota and their relationship with health and disease. However, our understanding of fish intestinal microbiota is limited. This review provides an overview of research on fish gut microbiota, including microbial composition, formation, factors that affect the GI microbes and characteristics of fish intestinal microbiota compared with human and mice. Further, the updated research on gnotobiotic zebrafish is elaborated and the insights gained on functions of the fish intestinal microbiota are discussed. Understanding the intestinal microbiota of fish will guide the development of probiotics, prebiotics and hopefully probiotic effectors as novel additives to improve the health of fish.
Uncertain and potentially harsh operating environments are often known to alter the operational performance of a system. In order to maintain system performance while coping with varying operating ...environments and potential disruptions, the resilience of engineered systems is desirable. Engineering systems are often interconnected in a dimensional way inherently from basic components to subsystems to the system of systems, which poses a grand challenge for system designers to analyze the resilience of such a complex system. Moreover, further complications in the assessment of resilience in the engineering domain are attributed to time-varying system performances, random perturbation occurrences, and probable failures caused by adverse events. This paper presents a dynamic Bayesian network (DBN) approach for the modeling and predictive resilience analysis for dynamic engineered systems. With the inter-time-slice links and the conditional probability tables in a DBN, the system performance could be molded as changing in a discrete time slice, while capturing the temporal probabilistic dependencies between the variables. An industrial-based case study of an electricity distribution system is further studied to demonstrate the effectiveness of the DBN approach for resilience analysis. The approach presented in this paper hopes to aid in realizing resiliency in system designs and to pave the way toward enhancements in developing resilient engineered systems.
In many real-world applications, data are often collected in the form of a stream, and thus the distribution usually changes in nature, which is referred to as
concept drift
in the literature. We ...propose a novel and effective approach to handle concept drift via model reuse, that is, reusing models trained on previous data to tackle the changes. Each model is associated with a weight representing its reusability towards current data, and the weight is adaptively adjusted according to the performance of the model. We provide both generalization and regret analysis to justify the superiority of our approach. Experimental results also validate its efficacy on both synthetic and real-world datasets.
Current machine learning techniques have achieved great success; however, there are many deficiencies. First, to train a strong model, a large amount of training examples are required, whereas ...collecting the data, particularly data with labels, is expensive or even difficult in many real tasks. Second, once a model has been trained, if environment changes, which often happens in real tasks, the model can hardly perform well or even become useless. Third,
Multi-Label Learning with Emerging New Labels Zhu, Yue; Ting, Kai Ming; Zhou, Zhi-Hua
IEEE transactions on knowledge and data engineering,
10/2018, Volume:
30, Issue:
10
Journal Article
Peer reviewed
Open access
In a multi-label learning task, an object possesses multiple concepts where each concept is represented by a class label. Previous studies on multi-label learning have focused on a fixed set of class ...labels, i.e., the class label set of test data is the same as that in the training set. In many applications, however, the environment is dynamic and new concepts may emerge in a data stream. In order to maintain a good predictive performance in this environment, a multi-label learning method must have the ability to detect and classify instances with emerging new labels. To this end, we propose a new approach called Multi-label learning with Emerging New Labels (MuENL). It has three functions: classify instances on currently known labels, detect the emergence of a new label, and construct a new classifier for each new label that works collaboratively with the classifier for known labels. In addition, we show that MuENL can be easily extended to handle sparse high dimensional data streams by simply reducing the original dimensionality, and then applying MuENL on the reduced dimensional space. Our empirical evaluation shows the effectiveness of MuENL on several benchmark datasets and MuENLHD on the sparse high dimensional Weibo dataset.