This paper proposes a scaled model to investigate the dynamic characteristics of suspension string and LMS adaptive control of hoop flexible structure. Firstly, the lateral vibration equation of micro segment string i...
详细信息
algorithms for full-information online learning are classically tuned to minimize their worst-case regret. Modern algorithms additionally provide tighter guarantees outside the adversarial regime, most notably in the ...
详细信息
ISBN:
(纸本)9781713871088
algorithms for full-information online learning are classically tuned to minimize their worst-case regret. Modern algorithms additionally provide tighter guarantees outside the adversarial regime, most notably in the form of constant pseudoregret bounds under statistical margin assumptions. We investigate the multiscale extension of the problem where the loss ranges of the experts are vastly different. Here, the regret with respect to each expert needs to scale with its range, instead of the maximum overall range. We develop new multiscale algorithms, tuning schemes and analysis techniques to show that worst-case robustness and adaptation to easy data can be combined at a negligible cost. We further develop an extension with optimism and apply it to solve multiscale two-player zero-sum games. We demonstrate experimentally the superior performance of our scale-adaptive algorithm and discuss the subtle relationship of our results to Freund's 2016 open problem.
The prompt-based learning paradigm has gained much research attention recently. It has achieved state-of-the-art performance on several NLP tasks, especially in the few-shot scenarios. While steering the downstream ta...
详细信息
ISBN:
(纸本)9781713871088
The prompt-based learning paradigm has gained much research attention recently. It has achieved state-of-the-art performance on several NLP tasks, especially in the few-shot scenarios. While steering the downstream tasks, few works have been reported to investigate the security problems of the prompt-based models. In this paper, we conduct the first study on the vulnerability of the continuous prompt learning algorithm to backdoor attacks. We observe that the few-shot scenarios have posed a great challenge to backdoor attacks on the prompt-based models, limiting the usability of existing NLP backdoor methods. To address this challenge, we propose BadPrompt, a lightweight and task-adaptive algorithm, to backdoor attack continuous prompts. Specially, BadPrompt first generates candidate triggers which are indicative for predicting the targeted label and dissimilar to the samples of the non-targeted labels. Then, it automatically selects the most effective and invisible trigger for each sample with an adaptive trigger optimization algorithm. We evaluate the performance of BadPrompt on five datasets and two continuous prompt models. The results exhibit the abilities of BadPrompt to effectively attack continuous prompts while maintaining high performance on the clean test sets, outperforming the baseline models by a large margin. The source code of BadPrompt is publicly available (1).
For high control performance of cable-driven manipulators, we design a new adaptive time-delay control (ATDC) using enhanced nonsingular fast terminal sliding mode (NFTSM). The proposed ATDC uses time-delay estimation...
详细信息
For high control performance of cable-driven manipulators, we design a new adaptive time-delay control (ATDC) using enhanced nonsingular fast terminal sliding mode (NFTSM). The proposed ATDC uses time-delay estimation (TDE) to acquire the lumped dynamics in a simple way and founds a practical model-free structure. Then, a new enhanced NFTSM surface is developed to ensure fast convergence and high control accuracy. To acquire good comprehensive performance under lumped uncertainties, in this article we propose a novel adaptive algorithm for the control gain, which can regulate itself based on the control errors timely and accurately. Benefitting from the TDE and the proposed enhanced NFTSM surface and adaptive control gain, our proposed ATDC is model-free, fast response, and accurate. Theoretical analysis concerning system stability, and control precision and convergence speed are given based on Lyapunov theory. Finally, the advantages of our ATDC over existing methods are verified with comparative experiments.
In solving simulation-based stochastic root-finding or optimization problems that involve rare events, such as in extreme quantile estimation, running crude Monte Carlo can be prohibitively inefficient. To address thi...
详细信息
In solving simulation-based stochastic root-finding or optimization problems that involve rare events, such as in extreme quantile estimation, running crude Monte Carlo can be prohibitively inefficient. To address this issue, importance sampling can be employed to drive down the sampling error to a desirable level. However, selecting a good importance sampler requires knowledge of the solution to the problem at hand, which is the goal to begin with and thus forms a circular challenge. We investigate the use of adaptive importance sampling to untie this circularity. Our procedure sequentially updates the importance sampler to reach the optimal sampler and the optimal solution simultaneously, and can be embedded in both sample-average-approximation-type algorithms and stochastic-approximation-type algorithms. Our theoretical analysis establishes strong consistency and asymptotic normality of the resulting estimators. We also demonstrate, via a minimax perspective, the key role of using adaptivity in controlling asymptotic errors. Finally, we illustrate the effectiveness of our approach via numerical experiments.
We study the problem of estimating the value of a known smooth function f at an unknown point mu is an element of R-n, where each component mu(i) can be sampled via a noisy oracle. Sampling more frequently components ...
详细信息
We study the problem of estimating the value of a known smooth function f at an unknown point mu is an element of R-n, where each component mu(i) can be sampled via a noisy oracle. Sampling more frequently components of mu corresponding to directions of the function with larger directional derivatives is more sample-efficient. However, as mu is unknown, the optimal sampling frequencies are also unknown. We design an instance-adaptive algorithm that learns to sample according to the importance of each coordinate, and with probability at least 1 - delta returns an epsilon accurate estimate of f(mu). We generalize our algorithm to adapt to heteroskedastic noise, and prove asymptotic optimality when f is linear. We corroborate our theoretical results with numerical experiments, showing the dramatic gains afforded by adaptivity.
This article presents recent outcomes of the author's research on musical complex adaptive systems (CASs). The first part focuses on the concepts of adaptation and complexity within the framework of CASs and sugge...
详细信息
This article presents recent outcomes of the author's research on musical complex adaptive systems (CASs). The first part focuses on the concepts of adaptation and complexity within the framework of CASs and suggests a rigorous placing of the concepts within the musical domain. This analysis involves a distinction of the notions of context and information between the engineering field of information theory and the philosophical one of radical constructivism. I conclude this section by showing that, in this approach, information and context are mutually determining. Then, I introduce a technique related to the notion of evolvability in biology and genetic algorithms and that has significantly increased the complexity and long-term variety in music systems during autonomous evolutions. This technique distributes adaptation across higher levels and allows the system to reorganize the relationships among its agents and their structure circularly while interpreting and constructing its context. To conclude, an autonomous live performance piece from 2019-2020, "Constructing Realities (Homage to Heinz von Foerster)," which implements the theories mentioned above, is described, showing DSP processes and techniques that relate to evolvability, autopoiesis, fitness, and complexity through agent-based modeling. This article is accompanied by a companion article discussing the technical aspects of information processing algorithms, which are an essential part for the implementation of music CASs: "Time-Domain adaptive algorithms for Low- and High-Level Audio Information Processing."
This study proposes AVARA, an innovative adaptive Voronoi diagram algorithm explicitly designed for residential areas containing adjacent polygons. AVARA addresses the complex task of balancing computational efficienc...
详细信息
This study proposes AVARA, an innovative adaptive Voronoi diagram algorithm explicitly designed for residential areas containing adjacent polygons. AVARA addresses the complex task of balancing computational efficiency and boundary position accuracy in existing Voronoi algorithms based on occurrence element discretization and difficulties constructing Voronoi diagrams for topologically adjacent polygonal objects. AVARA leverages interpolation strategies established on the classification result of neighbor pairs using co-edge relationships of residential area centroids. Additionally, AVARA incorporates graphical Boolean operations and node merging. The experiments used residential area datasets at 1:50,000, 1:10,000, and 1:2,000 map scales. The results showed that AVARA outperformed the Voronoi diagram algorithms based on intervisible points (VBIVP) and the equal-interval dense point method (VBEDP). AVARA achieved the highest accuracy rate in constructing Voronoi diagrams, with a 50% improvement. It also exhibited superior local position accuracy, demonstrating significant time performance improvements of 349.12% and 90.7% using datasets at map scales of 1:50,000 and 1:10,000, respectively. In the experiment using a dataset at a map scale of 1:2,000, AVARA consumed slightly less time than VBEDP with a 0.6 m interval. These findings highlight the significant overall advantage of AVARA. Additionally, the effectiveness of the algorithm was further validated using a large-scale building dataset.
In this work, a new class of stochastic gradient algorithm is developed based on fractional calculus. Unlike the existing algorithms, the concept of complex fractional gradient is introduced by employing Caputo's ...
详细信息
In this work, a new class of stochastic gradient algorithm is developed based on fractional calculus. Unlike the existing algorithms, the concept of complex fractional gradient is introduced by employing Caputo's fractional derivative which results in a fractional steepest descent algorithm and a fractional-order complex LMS (FoCLMS) algorithm. We demonstrate that with the Caputo's fractional gradient definition, the Weiner solution remains invariant. Convergence analysis of the proposed FoCLMS algorithm is presented for both transient and steady state scenarios. Consequently, expressions for the learning curves and steady state EMSE are derived. Our theoretical developments are validated by simulation experiments. Extensive simulations are presented to investigate all possible scenarios: channel with negative weights and real input data, channel with positive weights and complex input data, and channel with complex weights and complex input data.
This article compiles several aspects of the dynamics of stochastic approximation algorithms with Markov iterate-dependent noise when the iterates are not known to be stable beforehand. We achieve the same by extendin...
详细信息
This article compiles several aspects of the dynamics of stochastic approximation algorithms with Markov iterate-dependent noise when the iterates are not known to be stable beforehand. We achieve the same by extending the lock-in probability (i.e., the probability of convergence of the iterates to a specific attractor of the limiting ordinary differential equation (o.d.e.) given that the iterates are in its domain of attraction after a sufficiently large number of iterations (say) n(0)) framework to such recursions. Specifically, with the more restrictive assumption of Markov iterate-dependent noise supported on a bounded subset of the Euclidean space, we give a lower bound for the lock-in probability. We use these results to prove almost sure convergence of the iterates to the specified attractor when the iterates satisfy an asymptotic tightness condition. The novelty of our approach is that if the state space of the Markov process is compact, we prove almost sure convergence under much weaker assumptions compared to the work by Andrieu et al., which solves the general state-space case under much restrictive assumptions by providing sufficient conditions for stability of the iterates. We also extend our single-timescale results to the case where there are two separate recursions over two different timescales. This, in turn, is shown to be useful in analyzing the tracking ability of general adaptive algorithms. Additionally, we show that our results can be used to derive a sample complexity estimate of such recursions, which then can be used for step-size selection.
暂无评论