scholarly journals A glass-box interactive machine learning approach for solving NP-hard problems with the human-in-the-loop

2019 ◽  
Vol 28 (2) ◽  
pp. 121-134 ◽  
Author(s):  
ANDREAS HOLZINGER ◽  
MARKUS PLASS ◽  
KATHARINA HOLZINGER ◽  
GLORIA CERASELA CRIS¸AN ◽  
CAMELIA-M. PINTEA ◽  
...  
Keyword(s):  
Machine Learning ◽  
Protein Folding ◽  
Health Informatics ◽  
Learning Algorithm ◽  
Black Box ◽  
Np Hard ◽  
Interactive Machine Learning ◽  
Human In The Loop ◽  
Hard Problems ◽  
Np Hard Problems

The ultimate goal of the Machine Learning (ML) community is to develop algorithms that can automatically learn from data, to extract knowledge and to make decisions without any human intervention. Specifically, automatic Machine Learning (aML) approaches show impressive success, e.g. in speech/image recognition or autonomous drive and smart car industry. Recent results even demonstrate intriguingly that deep learning applied for automatic classification of skin lesions is on par with the performance of dermatologists, yet outperforms the average human efficiency. As human perception is inherently limited to 3D environments, such approaches can discover patterns, e.g. that two objects are similar, in arbitrarily high-dimensional spaces what no human is able to do. Humans can deal simultaneously only with limited amounts of data, whilst “big data” is not only beneficial but necessary for aML. However, in health informatics, there are few data sets; aML approaches often suffer from insufficient training samples. Many problems are computationally hard, e.g. subspace clustering, k-anonymization, or protein folding. Here, interactive machine learning (iML) could be successfully used, as a human-in-the-loop contributes to reduce a huge search space through heuristic selection of suitable samples. This can reduce the complexity of NP-hard problems through the knowledge brought in by a human agent involved into the learning algorithm. A huge motivation for iML is that standard black-box approaches lack transparency, hence do not foster trust and acceptance of ML among end-users. Most of all, rising legal and privacy aspects, e.g. the European General Data Protection Regulations (GDPR) make black-box approaches difficult to use, because they often are not able to explain why a decision has been made, e.g. why two objects are similar. All these reasons motivate the idea to open the black-box to a glass-box. In this paper, we present some experiments to demonstrate the effectiveness of the iML human-in-the-loop model, in particular when using a glass-box instead of a black-box model and thus enabling a human directly to interact with a learning algorithm. We selected the Ant Colony System (ACS) algorithm, and applied it on the Traveling Salesman Problem (TSP). The TSP-problem is a good example, because it is of high relevance for health informatics as for example on protein folding problem, thus of enormous importance for fostering cancer research. Finally, from studies of learning from observation, i.e. of how humans extract so much from so little data, fundamental ML-research also may benefit.

A note on no-free-lunch theorem

2021 ◽  
Author(s):  
Lidong Wu
Keyword(s):  
Machine Learning ◽  
Theoretical Result ◽  
Data Driven ◽  
Np Hard ◽  
Free Lunch ◽  
No Free Lunch Theorem ◽  
Data Driven Approach ◽  
Hard Problems ◽  
No Free Lunch ◽  
Np Hard Problems

The No-Free-Lunch theorem is an interesting and important theoretical result in machine learning. Based on philosophy of No-Free-Lunch theorem, we discuss extensively on the limitation of a data-driven approach in solving NP-hard problems.

Characteristics that Influence Perceived Intelligence in AI Design

2018 ◽  
Vol 62 (1) ◽  
pp. 1637-1641 ◽  
Author(s):  
Samantha Krening ◽  
Karen M. Feigh
Keyword(s):  
Machine Learning ◽  
Teaching Methods ◽  
Intelligent Agents ◽  
Learning Algorithm ◽  
Language Instruction ◽  
Design Decisions ◽  
Interactive Machine Learning ◽  
Human In The Loop ◽  
Learning Agent ◽  
Design Characteristics

A goal of interactive machine learning (IML) is to create robots or intelligent agents that can be easily taught how to perform tasks by individuals with no specialized training. To achieve that goal, researchers and designers must understand how certain design decisions impact the human’s experience of teaching the agent, such as influencing the agent’s perceived intelligence. We posit that the type of feedback a robot can learn from affects the perceived intelligence of the robot, similar to its physical appearance. This study investigated two methods of natural language instruction: critique and action advice. We conducted a human-in-the-loop experiment in which people trained two agents with different teaching methods but, unknown to each participant, the same underlying machine learning algorithm. The results show an agent that learns from binary good/bad critique is perceived as less intelligent than an agent that can learn from action instructions, even if the underlying machine learning agent is the same. In addition to the complexity of the input, other design characteristics we found that influence the agent’s perceived intelligence are: compliance, responsiveness, effort, transparency, and robustness.

scholarly journals Fifty years of P vs. NP and the possibility of the impossible

2022 ◽  
Vol 65 (1) ◽  
pp. 76-85
Author(s):  
Lance Fortnow
Keyword(s):  
Machine Learning ◽  
Np Hard ◽  
Hard Problems ◽  
Np Hard Problems

Advances in algorithms, machine learning, and hardware can help tackle many NP-hard problems once thought impossible.

scholarly journals Using Machine Learning for Quantum Annealing Accuracy Prediction

Algorithms ◽  
2021 ◽  
Vol 14 (6) ◽  
pp. 187
Author(s):  
Aaron Barbosa ◽  
Elijah Pelofske ◽  
Georg Hahn ◽  
Hristo N. Djidjev
Keyword(s):  
Machine Learning ◽  
Maximum Clique ◽  
Classification Model ◽  
Maximum Clique Problem ◽  
Problem Instance ◽  
Np Hard ◽  
Machine Learning Classification ◽  
Hard Problems ◽  
Problem Instances ◽  
D Wave

Quantum annealers, such as the device built by D-Wave Systems, Inc., offer a way to compute solutions of NP-hard problems that can be expressed in Ising or quadratic unconstrained binary optimization (QUBO) form. Although such solutions are typically of very high quality, problem instances are usually not solved to optimality due to imperfections of the current generations quantum annealers. In this contribution, we aim to understand some of the factors contributing to the hardness of a problem instance, and to use machine learning models to predict the accuracy of the D-Wave 2000Q annealer for solving specific problems. We focus on the maximum clique problem, a classic NP-hard problem with important applications in network analysis, bioinformatics, and computational chemistry. By training a machine learning classification model on basic problem characteristics such as the number of edges in the graph, or annealing parameters, such as the D-Wave’s chain strength, we are able to rank certain features in the order of their contribution to the solution hardness, and present a simple decision tree which allows to predict whether a problem will be solvable to optimality with the D-Wave 2000Q. We extend these results by training a machine learning regression model that predicts the clique size found by D-Wave.

scholarly journals A Review on Human–AI Interaction in Machine Learning and Insights for Medical Applications

2021 ◽  
Vol 18 (4) ◽  
pp. 2121
Author(s):  
Mansoureh Maadi ◽  
Hadi Akbarzadeh Khorshidi ◽  
Uwe Aickelin
Keyword(s):  
Machine Learning ◽  
Future Research ◽  
Computational Power ◽  
Medical Field ◽  
Interactive Machine Learning ◽  
Human In The Loop ◽  
Human Interactions ◽  
Scoping Literature Review ◽  
Domain Expertise ◽  
Expertise Level

Objective: To provide a human–Artificial Intelligence (AI) interaction review for Machine Learning (ML) applications to inform how to best combine both human domain expertise and computational power of ML methods. The review focuses on the medical field, as the medical ML application literature highlights a special necessity of medical experts collaborating with ML approaches. Methods: A scoping literature review is performed on Scopus and Google Scholar using the terms “human in the loop”, “human in the loop machine learning”, and “interactive machine learning”. Peer-reviewed papers published from 2015 to 2020 are included in our review. Results: We design four questions to investigate and describe human–AI interaction in ML applications. These questions are “Why should humans be in the loop?”, “Where does human–AI interaction occur in the ML processes?”, “Who are the humans in the loop?”, and “How do humans interact with ML in Human-In-the-Loop ML (HILML)?”. To answer the first question, we describe three main reasons regarding the importance of human involvement in ML applications. To address the second question, human–AI interaction is investigated in three main algorithmic stages: 1. data producing and pre-processing; 2. ML modelling; and 3. ML evaluation and refinement. The importance of the expertise level of the humans in human–AI interaction is described to answer the third question. The number of human interactions in HILML is grouped into three categories to address the fourth question. We conclude the paper by offering a discussion on open opportunities for future research in HILML.

scholarly journals In some curved spaces, one can solve NP-hard problems in polynomial time

2009 ◽  
Vol 158 (5) ◽  
pp. 727-740 ◽  
Author(s):  
V. Kreinovich ◽  
M. Margenstern
Keyword(s):  
Polynomial Time ◽  
Np Hard ◽  
Curved Spaces ◽  
Hard Problems ◽  
Np Hard Problems

NP vs QMA_\log(2)

2010 ◽  
Vol 10 (1&2) ◽  
pp. 141-151
Author(s):  
S. Beigi
Keyword(s):  
Quantum Algorithms ◽  
Np Hard ◽  
Complete Problems ◽  
Hard Problems ◽  
Np Complete ◽  
Np Hard Problems

Although it is believed unlikely that $\NP$-hard problems admit efficient quantum algorithms, it has been shown that a quantum verifier can solve NP-complete problems given a "short" quantum proof; more precisely, NP\subseteq QMA_{\log}(2) where QMA_{\log}(2) denotes the class of quantum Merlin-Arthur games in which there are two unentangled provers who send two logarithmic size quantum witnesses to the verifier. The inclusion NP\subseteq QMA_{\log}(2) has been proved by Blier and Tapp by stating a quantum Merlin-Arthur protocol for 3-coloring with perfect completeness and gap 1/24n^6. Moreover, Aaronson et al. have shown the above inclusion with a constant gap by considering $\widetilde{O}(\sqrt{n})$ witnesses of logarithmic size. However, we still do not know if QMA_{\log}(2) with a constant gap contains NP. In this paper, we show that 3-SAT admits a QMA_{\log}(2) protocol with the gap 1/n^{3+\epsilon}} for every constant \epsilon>0.

scholarly journals Coupling Elephant Herding with Ordinal Optimization for Solving the Stochastic Inequality Constrained Optimization Problems

2020 ◽  
Vol 10 (6) ◽  
pp. 2075 ◽  
Author(s):  
Shih-Cheng Horng ◽  
Shieh-Shing Lin
Keyword(s):  
Optimization Problems ◽  
Solution Space ◽  
Inequality Constraints ◽  
Optimization Methods ◽  
Ordinal Optimization ◽  
Np Hard ◽  
Constrained Optimization Problems ◽  
Inequality Constrained Optimization ◽  
Hard Problems ◽  
Np Hard Problems

The stochastic inequality constrained optimization problems (SICOPs) consider the problems of optimizing an objective function involving stochastic inequality constraints. The SICOPs belong to a category of NP-hard problems in terms of computational complexity. The ordinal optimization (OO) method offers an efficient framework for solving NP-hard problems. Even though the OO method is helpful to solve NP-hard problems, the stochastic inequality constraints will drastically reduce the efficiency and competitiveness. In this paper, a heuristic method coupling elephant herding optimization (EHO) with ordinal optimization (OO), abbreviated as EHOO, is presented to solve the SICOPs with large solution space. The EHOO approach has three parts, which are metamodel construction, diversification and intensification. First, the regularized minimal-energy tensor-product splines is adopted as a metamodel to approximately evaluate fitness of a solution. Next, an improved elephant herding optimization is developed to find N significant solutions from the entire solution space. Finally, an accelerated optimal computing budget allocation is utilized to select a superb solution from the N significant solutions. The EHOO approach is tested on a one-period multi-skill call center for minimizing the staffing cost, which is formulated as a SICOP. Simulation results obtained by the EHOO are compared with three optimization methods. Experimental results demonstrate that the EHOO approach obtains a superb solution of higher quality as well as a higher computational efficiency than three optimization methods.

Sign in / Sign up

Export Citation Format

Share Document