Design Of Learning Algorithms Research Articles

Error Density-dependent Empirical Risk Minimization

Empirical Risk Minimization (ERM) with the squared loss has become one of the most popular principles for designing learning algorithms. However, the existing mean regression models under ERM usually suffer from poor generalization performance due to their sensitivity to atypical observations (e.g., outliers). For alleviating this sensitivity, some strategies have been proposed by utilizing the quantitative relationship of error values with different observations to form robust learning objectives. Instead of focusing on error values, this paper considers the error density to uncover the structure information of observations and proposes a new learning objective, called Error Density-dependent Empirical Risk Minimization (EDERM), for robust regression under complex data environments. Property characterizations and experimental analysis validate the robustness and competitiveness of the proposed EDERM-based learning models. The implemented codes can be found at https://github.com/zhangxuelincode/EDERM.

Designing Biological Sequences without Prior Knowledge Using Evolutionary Reinforcement Learning

Designing novel biological sequences with desired properties is a significant challenge in biological science because of the extra large search space. The traditional design process usually involves multiple rounds of costly wet lab evaluations. To reduce the need for expensive wet lab experiments, machine learning methods are used to aid in designing biological sequences. However, the limited availability of biological sequences with known properties hinders the training of machine learning models, significantly restricting their applicability and performance. To fill this gap, we present ERLBioSeq, an Evolutionary Reinforcement Learning algorithm for BIOlogical SEQuence design. ERLBioSeq leverages the capability of reinforcement learning to learn without prior knowledge and the potential of evolutionary algorithms to enhance the exploration of reinforcement learning in the large search space of biological sequences. Additionally, to enhance the efficiency of biological sequence design, we developed a predictor for sequence screening in the biological sequence design process, which incorporates both the local and global sequence information. We evaluated the proposed method on three main types of biological sequence design tasks, including the design of DNA, RNA, and protein. The results demonstrate that the proposed method achieves significant improvement compared to the existing state-of-the-art methods.

Improved learning in human evolutionary systems with dynamic contrastive learning

We introduce a new inductive bias for learning in dynamic event-based human systems. This is intended to partially address the issue of deep learning in chaotic systems. Instead of fitting the data to polynomial expansions that are expressive enough to approximate the generative functions or of inducing a universal approximator to learn the patterns and inductive bias, we only assume that the relationship between the input features and output classes changes over time, and embed this assumption through a form of dynamic contrastive learning in pre-training, where pre-training labels contain information about the class labels and time periods. We do this by extending and integrating two separate forms of contrastive learning. We note that this approach is not equivalent to inserting an extra feature into the input data that contains time period, because the input data cannot contain the label. We illustrate the approach on a recently designed learning algorithm for event-based graph time-series classification, and demonstrate its value on real-world data.

A multiphase study protocol of identifying, and predicting cancer-related symptom clusters: applying a mixed-method design and machine learning algorithms.

In recent years, there has been increasing attention on the cluster approach to symptom management. Two significant challenges in the symptom cluster (SC) approach are identifying and predicting these clusters. This multiphase protocol aims to identify SCs in patients with advanced cancer as the primary objective, with the secondary objective of developing machine learning algorithms to predict SCs identified in the first phase. The 2-MIXIP study consists of two main phases. The first phase involves identifying SCs, and the second phase focuses on developing predictive algorithms for the identified SCs. The identification of SCs involves a parallel mixed-method design (quantitative and qualitative). Quantitative and qualitative methods are conducted simultaneously and given equal importance. The data are collected and analyzed independently before being integrated. The quantitative part is conducted using a descriptive-analytical method. The qualitative analysis is conducted using a content analysis approach. Then, the identified SCs from both parts are integrated to determine the final clusters and use them in the second phase. In the second phase, we employ a tree-based machine learning method to create predictive algorithms for SCs using key demographic and clinical patient characteristics. The findings of the 2-MIXIP study can help manage cancer patients' symptoms more effectively and enhance clinical decision-making by using SCs prediction. Furthermore, the results of this study can provide guidance for clinical trials aimed at managing symptoms.

Retracted: Fuzzy Logic-Based Machine Learning Algorithm for Cultural and Creative Product Design.

[This retracts the article DOI: 10.1155/2022/7747192.].

A parallel heterogeneous policy deep reinforcement learning algorithm for bipedal walking motion design.

Considering the dynamics and non-linear characteristics of biped robots, gait optimization is an extremely challenging task. To tackle this issue, a parallel heterogeneous policy Deep Reinforcement Learning (DRL) algorithm for gait optimization is proposed. Firstly, the Deep Deterministic Policy Gradient (DDPG) algorithm is used as the main architecture to run multiple biped robots in parallel to interact with the environment. And the network is shared to improve the training efficiency. Furthermore, heterogeneous experience replay is employed instead of the traditional experience replay mechanism to optimize the utilization of experience. Secondly, according to the walking characteristics of biped robots, a biped robot periodic gait is designed with reference to sinusoidal curves. The periodic gait takes into account the effects of foot lift height, walking period, foot lift speed and ground contact force of the biped robot. Finally, different environments and different biped robot models pose challenges for different optimization algorithms. Thus, a unified gait optimization framework for biped robots based on the RoboCup3D platform is established. Comparative experiments were conducted using the unified gait optimization framework, and the experimental results show that the method outlined in this paper can make the biped robot walk faster and more stably.

Modeling of Hybrid Henry Gas Solubility Optimization Algorithm with Deep Learning-Based LED Driver System

Light emitting diodes (LEDs) have become an effective lighting solution because of the characteristics of energy efficiency, flexible controllability and extended lifetime. They find use in numerous lighting systems for residents, industries, enterprises and street lighting applications. The efficiency and trustworthiness of the LED systems considerably based on the thermal mechanical loading improved several degradation schemes and respective interfaces. The complication of the LED systems limits the theoretic interpretation of the core reasons for the luminous variation or the formation of the direct correlation among the thermal aging loading and the luminous output. Therefore, this paper designs a new hybrid Henry gas solubility optimization with deep learning (HHGSO-DL) algorithm for LED driver system design. The presented HHGSO-DL technique mainly concentrates on the derivation of empirical relationships among the design parameters, thermal aging loading and luminous outcomes of the LED product. In the presented HHGSO-DL technique, bidirectional long short-term memory (BiLSTM) algorithm is executed for examining the empirical relationship and its hyperparameters can be tuned by the HHGSO algorithm. In this work, the HHGSO algorithm is derived by the integration of traditional HGSO algorithm with oppositional-based learning (OBL) concept. The performance of the HHGSO-DL technique can be investigated on LED chip packaging and LED luminaire with thermal aging loading. The extensive results demonstrate the promising performance of the HHGSO-DL technique over other state-of-the-art approaches.

Optimizing the efficient removal of ciprofloxacin from aqueous solutions by polyethylene terephthalate microplastics using multivariate statistical approach

Emerging pollutants such as ciprofloxacin (CIP) and microplastics (MPs) has been found in environment, as a result of their widespread usage. In this study, the use of different polyethylene terephthalate microplastics (PET MPs) such as pristine (Pr-PET), acid modified (Mod-PET) and thermal-oxidatively aged (Ag-PET) were optimized for their removal efficiencies for CIP (5 to 50 mg/L) from aqueous solution using Response Surface Methodology with Central Composite Design (4 factors − 2 levels) and Machine Learning (ML) algorithm such as Artificial Neural Networks (ANN), Decision Tree (DT), Random Forest (RF), and Support Vector Machine (SVM). Through the use of the desirability function, the optimal conditionswere determined and amaximum removal efficiency of93.95%, 94.36%, and 96.29% for Pr-PET, Mod-PET, and Ag-PET MPs, respectively. The Dumwald-Wagner intraparticle model and the Dual-Exponential Model (DEM) were used to describe the adsorption process. The Dumwald-Wagner modelwas able to describe the adsorption process before equilibrium was reached at 12 hrs. The DEM model divided the adsorption process into two stages, a rapid, initial stage and a slower, secondary stage.These results suggest that the use of PET MPs for removing CIP through adsorption is a viable option and can be implemented in treatment systems.

Design of experiments for the calibration of history-dependent models via deep reinforcement learning and an enhanced Kalman filter

Experimental data are often costly to obtain, which makes it difficult to calibrate complex models. For many models an experimental design that produces the best calibration given a limited experimental budget is not obvious. This paper introduces a deep reinforcement learning (RL) algorithm for design of experiments that maximizes the information gain measured by Kullback–Leibler divergence obtained via the Kalman filter (KF). This combination enables experimental design for rapid online experiments where manual trial-and-error is not feasible in the high-dimensional parametric design space. We formulate possible configurations of experiments as a decision tree and a Markov decision process, where a finite choice of actions is available at each incremental step. Once an action is taken, a variety of measurements are used to update the state of the experiment. This new data leads to a Bayesian update of the parameters by the KF, which is used to enhance the state representation. In contrast to the Nash–Sutcliffe efficiency index, which requires additional sampling to test hypotheses for forward predictions, the KF can lower the cost of experiments by directly estimating the values of new data acquired through additional actions. In this work our applications focus on mechanical testing of materials. Numerical experiments with complex, history-dependent models are used to verify the implementation and benchmark the performance of the RL-designed experiments.

Comparative study of lipid nanoparticle-based mRNA vaccine bioprocess with machine learning and combinatorial artificial neural network-design of experiment approach

To develop a combinatorial artificial-neural-network design-of-experiment (ANN-DOE) model, the effect of ionizable lipid, an ionizable lipid-to-cholesterol ratio, N/P ratio, flow rate ratio (FRR), and total flow rate (TFR) on the outcome responses of mRNA-LNP vaccine were evaluated using a definitive screening design (DSD) and machine learning (ML) algorithms. Particle size (PS), PDI, zeta potential (ZP), and encapsulation efficiency (EE) of mRNA-LNP were optimized within a defined constraint (PS 40–100 nm, PDI ≤ 0.30, ZP≥(±)0.30 mV, EE ≥ 70 %), fed to ML algorithms (XGBoost, bootstrap forest, support vector machines, k-nearest neighbors, generalized regression–Lasso, ANN) and prediction was compared to ANN-DOE model. Increased FRR decreased the PS and increased ZP, while increased TFR increased PDI and ZP. Similarly, DOTAP and DOTMA produced higher ZP and EE. Particularly, a cationic ionizable lipid with an N/P ratio ≥ 6 provided a higher EE. ANN showed better predictive ability (R2 = 0.7269–0.9946), while XGBoost demonstrated better RASE (0.2833–2.9817). The ANN-DOE model outperformed both optimized ML models by R2 = 1.21 % and RASE = 43.51 % (PS prediction), R2 = 0.23 % and RASE = 3.47 % (PDI prediction), R2 = 5.73 % and RASE = 27.95 % (ZP prediction), and R2 = 0.87 % and RASE = 36.95 % (EE prediction), respectively, which demonstrated that ANN-DOE model was superior in predicting the bioprocess compared to independent models.

MLPro-MPPS—A high-performance simulation framework for customizable production systems

Although the integration of machine learning into production systems has demonstrated significant potential, its real-life applications remain challenging. It is often necessary to rely on digital representations of the actual systems to design and test machine learning algorithms for controlling the systems because conducting such processes directly in real systems is expensive and risky. Hence, in this paper, we present a standardized, multi-purpose and flexible production systems simulator in Python for scientific and industrial activities, namely MLPro-MPPS, that is integrated with the MLPro package. Consequently, this allows the simulations by MLPro-MPPS to be compatible with machine learning tasks.

Exploring Neuromorphic Computing Based on Spiking Neural Networks: Algorithms to Hardware

Neuromorphic Computing, a concept pioneered in the late 1980s, is receiving a lot of attention lately due to its promise of reducing the computational energy, latency, as well as learning complexity in artificial neural networks. Taking inspiration from neuroscience, this interdisciplinary field performs a multi-stack optimization across devices, circuits, and algorithms by providing an end-to-end approach to achieving brain-like efficiency in machine intelligence. On one side, neuromorphic computing introduces a new algorithmic paradigm, known as Spiking Neural Networks (SNNs), which is a significant shift from standard deep learning and transmits information as spikes (“1” or “0”) rather than analog values. This has opened up novel algorithmic research directions to formulate methods to represent data in spike-trains, develop neuron models that can process information over time, design learning algorithms for event-driven dynamical systems, and engineer network architectures amenable to sparse, asynchronous, event-driven computing to achieve lower power consumption. On the other side, a parallel research thrust focuses on development of efficient computing platforms for new algorithms. Standard accelerators that are amenable to deep learning workloads are not particularly suitable to handle processing across multiple timesteps efficiently. To that effect, researchers have designed neuromorphic hardware that rely on event-driven sparse computations as well as efficient matrix operations. While most large-scale neuromorphic systems have been explored based on CMOS technology, recently, Non-Volatile Memory (NVM) technologies show promise toward implementing bio-mimetic functionalities on single devices. In this article, we outline several strides that neuromorphic computing based on spiking neural networks (SNNs) has taken over the recent past, and we present our outlook on the challenges that this field needs to overcome to make the bio-plausibility route a successful one.

Machine Learning Algorithm for Efficient Design of Separated Buffer Super-Junction IGBT

An improved structure for an Insulated Gate Bipolar Transistor (IGBT) with a separated buffer layer is presented in order to improve the trade-off between the turn-off loss (Eoff) and on-state voltage (Von). However, it is difficult to set efficient parameters due to the increase in the new buffer doping concentration variable. Therefore, a machine learning (ML) algorithm is proposed as a solution. Compared to the conventional Technology Computer-Aided Design (TCAD) simulation tool, it is demonstrated that incorporating the ML algorithm into the device analysis could make it possible to achieve high accuracy and significantly shorten the simulation time. Specifically, utilizing the ML algorithm could achieve coefficients of determination (R2) of Von and Eoff of 0.995 and 0.968, respectively. In addition, it enables the optimized design to fit the target characteristics. In this study, the structure proposed for the trade-off improvement was targeted to obtain the minimum Eoff at the same Von, especially by adjusting the concentration of the separated buffer. We could improve Eoff by 36.2% by optimizing the structure, which was expected to be improved by 24.7% using the ML approach. In another way, it is possible to inversely design four types of structures with characteristics close to the target characteristics (Eoff = 1.64 μJ, Von = 1.38 V). The proposed method of incorporating machine learning into device analysis is expected to be very strategic, especially for power electronics analysis (where the transistor size is comparatively large and requires significant computation). In summary, we improved the trade-off using a separated buffer, and ML enabled optimization and a more precise design, as well as reverse engineering.

Cooperative adaptive management of the Nile River with climate and socio-economic uncertainties

The uncertainties around the hydrological and socio-economic implications of climate change pose a challenge for Nile River system management, especially with rapidly rising demands for river-system-related services and political tensions between the riparian countries. Cooperative adaptive management of the Nile can help alleviate some of these stressors and tensions. Here we present a planning framework for adaptive management of the Nile infrastructure system, combining climate projections; hydrological, river system and economy-wide simulators; and artificial intelligence multi-objective design and machine learning algorithms. We demonstrate the utility of the framework by designing a cooperative adaptive management policy for the Grand Ethiopian Renaissance Dam that balances the transboundary economic and biophysical interests of Ethiopia, Sudan and Egypt. This shows that if the three countries compromise cooperatively and adaptively in managing the dam, the national-level economic and resilience benefits are substantial, especially under climate projections with the most extreme streamflow changes.

Flowsheet generation through hierarchical reinforcement learning and graph neural networks

AbstractProcess synthesis experiences a disruptive transformation accelerated by artificial intelligence. We propose a reinforcement learning algorithm for chemical process design based on a state‐of‐the‐art actor‐critic logic. Our proposed algorithm represents chemical processes as graphs and uses graph convolutional neural networks to learn from process graphs. In particular, the graph neural networks are implemented within the agent architecture to process the states and make decisions. We implement a hierarchical and hybrid decision‐making process to generate flowsheets, where unit operations are placed iteratively as discrete decisions and corresponding design variables are selected as continuous decisions. We demonstrate the potential of our method to design economically viable flowsheets in an illustrative case study comprising equilibrium reactions, azeotropic separation, and recycles. The results show quick learning in discrete, continuous, and hybrid action spaces. The method is predestined to include large action‐state spaces and an interface to process simulators in future research.

Fuzzy Logic-Based Machine Learning Algorithm for Cultural and Creative Product Design.

In order to effectively assist industrial designers in the color scheme design of cultural and creative products and output color schemes that meet users' image preferences, an interactive color scheme design method for cultural and creative products based on triangular fuzzy number is proposed. Through the group consistency decision based on triangular fuzzy number, an interactive genetic evolution of the color scheme population of cultural and creative products is performed to generate color schemes that satisfy the group consistency imagery preference and satisfaction, and the final selection scheme is calculated by color beauty. This paper verifies that the method can effectively integrate the imagery preferences of group users and help industrial designers to better design color schemes for cultural and creative products through the results of multiuser decision consistency.

Online-learning-aided optimization and interpretation of sugar production from oil palm mesocarp fibers with analytics for industrial applications

Conversion of oil palm mesocarp fibers (MF) into fermentable sugar through catalytic pretreatment coupled with enzymatic hydrolysis is a promising solution for biomass valorization. However, lack of understanding of the complicated conversion process and optimization of critical reaction conditions limits the efficacy of the sugar production (SP). In this study, we aim to develop an online framework incorporating orthogonal experimental design and machine learning (ML) algorithm to optimize important conditions and identify their interaction impacts on overall sugar yield from MF. The results show that both solid reduction (SR) and sugar conversion (SC) increase to 73.85% and 51.91% from ML-based optimization with relative errors of 20.58%±0.70% and 7.50%±3.68% from experimental validation, respectively, after two rounds of online learning. Model-based interpretation suggests that substrate and catalyst concentrations are two negatively related conditions with both SR and SC, while the solid concentration has a piecewise linear folded relationship. Moreover, potential implications on environments and economy of the presented efficiency (51.91%) of overall sugar yield were evaluated based on the amount of MF generated in Malaysia. The ML-optimized conversion strategy can save about 70 million tons of sugar cane and achieve a profit of US$ 1,450 million annually with a low operation cost (40 US$/ton MF conversion). It provides insights that help us to identify important conditions for improving conversion efficiency and promoting future industrialization.

Waiting But Not Aging: Optimizing Information Freshness Under the Pull Model

The Age-of-Information is an important metric for investigating the timeliness performance in information-update systems. In this paper, we study the AoI minimization problem under a new Pull model with replication schemes, where a user proactively sends a replicated request to multiple servers to “pull” the information of interest. Interestingly, we find that under this new Pull model, replication schemes capture a novel tradeoff between different values of the AoI across the servers (due to the random updating processes) and different response times across the servers, which can be exploited to minimize the expected AoI at the user's side. Specifically, assuming Poisson updating process for the servers and exponentially distributed response time, we derive a closed-form formula for computing the expected AoI and obtain the optimal number of responses to wait for to minimize the expected AoI. Then, we extend our analysis to the setting where the user aims to maximize the AoI-based utility, which represents the user's satisfaction level with respect to freshness of the received information. Furthermore, we consider a more realistic scenario where the user has no prior knowledge of the system. In this case, we reformulate the utility maximization problem as a stochastic Multi-Armed Bandit problem with side observations and leverage a special linear structure of side observations to design learning algorithms with improved performance guarantees. Finally, we conduct extensive simulations to elucidate our theoretical results and compare the performance of different algorithms. Our findings reveal that under the Pull model, waiting does not necessarily lead to aging; waiting for more than one response can often significantly reduce the AoI and improve the AoI-based utility in most scenarios.

A Review on Neural Turing Machine (NTM)

One of the significant objectives of artificial intelligence is to design learning algorithms that are executed on general-purpose computational machines inspired by the human brain. Neural Turing Machine (NTM) is a step towards realizing such a computational machine. In the literature, a variety of approaches have been presented for the NTM; however, there is no existing comprehensive survey and taxonomy for NTM methods. This article presents an overview of taxonomies characterizing the critical concepts of the NTM through a comprehensive survey on the related research activities. This in-depth analysis of taxonomies can provide researchers, designers, and application developers with a clear guideline to compare NTM methods. The taxonomy of machine learning, neural networks, and the Turing machine is introduced. The NTM is also inspected in terms of concepts, structure, implemented tasks, and related works. The article further presents research discussions and future challenges in this area.

MaxiMin Active Learning in Overparameterized Model Classes

Generating labeled training datasets has become a major bottleneck in Machine Learning (ML) pipelines. Active ML aims to address this issue by designing learning algorithms that automatically and adaptively select the most informative examples for labeling so that human time is not wasted labeling irrelevant, redundant, or trivial examples. This paper proposes a new approach to active ML with nonparametric or overparameterized models such as kernel methods and neural networks. In the context of binary classification, the new approach is shown to possess a variety of desirable properties that allow active learning algorithms to automatically and efficiently identify decision boundaries and data clusters.