Interpreting Model Predictions with Constrained Perturbation and Counterfactual Instances

2021 ◽  
Author(s):  
Jun-Peng Fang ◽  
Jun Zhou ◽  
Qing Cui ◽  
Cai-Zhi Tang ◽  
Long-Fei Li
Keyword(s):  
Financial Markets ◽  
State Of The Art ◽  
Industrial Applications ◽  
Importance Value ◽  
Art Interpretation ◽  
Critical Areas ◽  
Black Boxes ◽  
Model Predictions ◽  
Interpretation Method ◽  
Made In

In recent years, machine learning models have achieved magnificent success in many industrial applications, but most of them are black boxes. It is crucial to understand why such predictions are made in many critical areas such as medicine, financial markets, and auto driving. In this paper, we propose Coco, a novel interpretation method which can interpret any binary classifier by assigning each feature an importance value for a particular prediction. We first adopt MixUp method to generate reasonable perturbations, then apply these perturbations with constraints to obtain counterfactual instances and finally compute a comprehensive metric on these instances to estimate the importance of each feature. To demonstrate the effectiveness of Coco, we conduct extensive experiments on several datasets. The results show our method achieves better performance in identifying the most important features compared with the state-of-the-art interpretation methods, including Shap and Lime.

scholarly journals Adversarial-Learned Loss for Domain Adaptation

2020 ◽  
Vol 34 (04) ◽  
pp. 3521-3528
Author(s):  
Minghao Chen ◽  
Shuai Zhao ◽  
Haifeng Liu ◽  
Deng Cai
Keyword(s):  
Domain Adaptation ◽  
State Of The Art ◽  
Confusion Matrix ◽  
Ground Truth ◽  
Adversarial Learning ◽  
Label Method ◽  
Model Predictions ◽  
Novel Method ◽  
Remarkable Progress ◽  
Made In

Recently, remarkable progress has been made in learning transferable representation across domains. Previous works in domain adaptation are majorly based on two techniques: domain-adversarial learning and self-training. However, domain-adversarial learning only aligns feature distributions between domains but does not consider whether the target features are discriminative. On the other hand, self-training utilizes the model predictions to enhance the discrimination of target features, but it is unable to explicitly align domain distributions. In order to combine the strengths of these two methods, we propose a novel method called Adversarial-Learned Loss for Domain Adaptation (ALDA). We first analyze the pseudo-label method, a typical self-training method. Nevertheless, there is a gap between pseudo-labels and the ground truth, which can cause incorrect training. Thus we introduce the confusion matrix, which is learned through an adversarial manner in ALDA, to reduce the gap and align the feature distributions. Finally, a new loss function is auto-constructed from the learned confusion matrix, which serves as the loss for unlabeled target samples. Our ALDA outperforms state-of-the-art approaches in four standard domain adaptation datasets. Our code is available at https://github.com/ZJULearning/ALDA.

Biological Production of (S)-acetoin: A State-of-the-Art Review

2019 ◽  
Vol 19 (25) ◽  
pp. 2348-2356 ◽  
Author(s):  
Neng-Zhong Xie ◽  
Jian-Xiu Li ◽  
Ri-Bo Huang
Keyword(s):  
Side Effects ◽  
Chemical Synthesis ◽  
State Of The Art ◽  
Optically Active ◽  
Considerable Progress ◽  
Future Prospects ◽  
Chiral Compound ◽  
Biological Production ◽  
Acetoin Production ◽  
Made In

Acetoin is an important four-carbon compound that has many applications in foods, chemical synthesis, cosmetics, cigarettes, soaps, and detergents. Its stereoisomer (S)-acetoin, a high-value chiral compound, can also be used to synthesize optically active drugs, which could enhance targeting properties and reduce side effects. Recently, considerable progress has been made in the development of biotechnological routes for (S)-acetoin production. In this review, various strategies for biological (S)- acetoin production are summarized, and their constraints and possible solutions are described. Furthermore, future prospects of biological production of (S)-acetoin are discussed.

scholarly journals Smoothing additive manufactured parts using ns-pulsed laser radiation

2021 ◽  
Author(s):  
Florian Kuisat ◽  
Fernando Lasagni ◽  
Andrés Fabián Lasagni
Keyword(s):  
Surface Roughness ◽  
Mechanical Performance ◽  
State Of The Art ◽  
Pulsed Laser ◽  
Industrial Applications ◽  
Laser Source ◽  
Pulsed Laser Radiation ◽  
Current State ◽  
The Impact

AbstractIt is well known that the surface topography of a part can affect its mechanical performance, which is typical in additive manufacturing. In this context, we report about the surface modification of additive manufactured components made of Titanium 64 (Ti64) and Scalmalloy®, using a pulsed laser, with the aim of reducing their surface roughness. In our experiments, a nanosecond-pulsed infrared laser source with variable pulse durations between 8 and 200 ns was applied. The impact of varying a large number of parameters on the surface quality of the smoothed areas was investigated. The results demonstrated a reduction of surface roughness Sa by more than 80% for Titanium 64 and by 65% for Scalmalloy® samples. This allows to extend the applicability of additive manufactured components beyond the current state of the art and break new ground for the application in various industrial applications such as in aerospace.

scholarly journals Absorption heat transformer - state-of-the-art of industrial applications

2021 ◽  
Vol 141 ◽  
pp. 110757 ◽  
Author(s):  
Falk Cudok ◽  
Niccolò Giannetti ◽  
José L. Corrales Ciganda ◽  
Jun Aoyama ◽  
P. Babu ◽  
...  
Keyword(s):  
State Of The Art ◽  
Industrial Applications ◽  
Heat Transformer

Developments in Rechargeable Mno2Electrodes for Lithium Batteries

1988 ◽  
Vol 135 ◽  
Author(s):  
Michael M Thackeray
Keyword(s):  
Lithium Batteries ◽  
Lithium Battery ◽  
Electrode Materials ◽  
State Of The Art ◽  
Structural Features ◽  
Rechargeable Batteries ◽  
Current State ◽  
Alternative Systems ◽  
Battery Technology ◽  
Made In

AbstractConsiderable efforts are in progress to develop rechargeable batteries as alternative systems to the nickel-cadmium battery. In this regard, several advances have been made in ambient-temperature lithium battery technology, and specifically in the engineering of rechargeable lithium/manganese dioxide cells. This paper reviews the current state of the art in rechargeable Li/MnO2battery technology; particular attention is paid to the structural features of various MnO2electrode materials which influence their electrochemical and cycling behaviour in lithium cells.

State of the Art — 1973

1973 ◽  
Vol 18 (3) ◽  
pp. 175-183 ◽  
Author(s):  
Nathan B. Epstein ◽  
Duane S. Bishop
Keyword(s):  
State Of The Art ◽  
Therapeutic Process ◽  
Scientific Discipline ◽  
Therapeutic Approaches ◽  
Training And Education ◽  
Rigorous Approach ◽  
The Family ◽  
Process And Outcome ◽  
Scientific Approach ◽  
Made In

In summary, it can be said that progress is being made in the field, but slowly. The ‘art’ is vigorous, vital and exciting. The ranks of family therapists are swelling and they are coming from backgrounds of different theoretical persuasions and with varying degrees of sophistication in their training and education. This mélange does lead to excitement and turbulence but often detracts from the necessary rigour that a scientific discipline must develop in order to reach maturity. Systems theory allows for easy conceptualization of one another's behaviour in the system, and permits a much clearer understanding of the therapeutic process based upon it, in contrast to therapeutic approaches based on other models. The authors found negotiation to be therapeutically effective when made explicit. In addition they place the focus on the ‘here and now’ and encourage the increased labeling by family members of interactions (affective and behavioural) and their effects (affective and behavioural), according to the Family Categories Schema previously referred to. Efforts are directed especially towards dealing with the current resistances to problem solutions. Epstein et al. have reported on an ongoing program of research which attempts to examine the process and outcome of family therapy (7,10,21,22,25,26). What is needed now is a more rigorous approach to research and the development of a necessary theoretical base in order that a more systematic and scientific approach can be developed for treating families.

scholarly journals Social and Solidarity Economy Web Information Systems

2014 ◽  
Vol 12 (1) ◽  
pp. 35-52 ◽  
Author(s):  
Mariana Curado Malta ◽  
Ana Alice Baptista ◽  
Cristina Parente
Keyword(s):  
Information Systems ◽  
State Of The Art ◽  
Bibliographic Databases ◽  
Sources Of Information ◽  
Solidarity Economy ◽  
Web Based ◽  
Web Information ◽  
The Social ◽  
Application Profile ◽  
Made In

This paper presents the state of the art on interoperability developments for the social and solidarity economy (SSE) community web based information systems (WIS); it also presents a framework of interoperability for the SSE' WIS and the developments made in a research-in-progress PhD project in the last 3 years. A search on the bibliographic databases showed that so far there are no papers on interoperability initiatives on the SSE, so it was necessary to have other sources of information: a preliminary analysis of the WIS that support SSE activities; and interviews with the representatives of some of the world's most important SSE organisations. The study showed that the WIS are still not interoperable yet. In order to become interoperable a group of the SSE community has been developing a Dublin Corre Application Profile to be used by the SSE community as reference and binding to describe their resources. This paper also describes this on-going process.

scholarly journals SetRank: A Setwise Bayesian Approach for Collaborative Ranking from Implicit Feedback

2020 ◽  
Vol 34 (04) ◽  
pp. 6127-6136
Author(s):  
Chao Wang ◽  
Hengshu Zhu ◽  
Chen Zhu ◽  
Chuan Qin ◽  
Hui Xiong
Keyword(s):  
Bayesian Approach ◽  
Posterior Probability ◽  
State Of The Art ◽  
User Preferences ◽  
Implicit Feedback ◽  
Pairwise Preference ◽  
Entire List ◽  
Collaborative Ranking ◽  
Real World Datasets ◽  
Made In

The recent development of online recommender systems has a focus on collaborative ranking from implicit feedback, such as user clicks and purchases. Different from explicit ratings, which reflect graded user preferences, the implicit feedback only generates positive and unobserved labels. While considerable efforts have been made in this direction, the well-known pairwise and listwise approaches have still been limited by various challenges. Specifically, for the pairwise approaches, the assumption of independent pairwise preference is not always held in practice. Also, the listwise approaches cannot efficiently accommodate “ties” due to the precondition of the entire list permutation. To this end, in this paper, we propose a novel setwise Bayesian approach for collaborative ranking, namely SetRank, to inherently accommodate the characteristics of implicit feedback in recommender system. Specifically, SetRank aims at maximizing the posterior probability of novel setwise preference comparisons and can be implemented with matrix factorization and neural networks. Meanwhile, we also present the theoretical analysis of SetRank to show that the bound of excess risk can be proportional to √M/N, where M and N are the numbers of items and users, respectively. Finally, extensive experiments on four real-world datasets clearly validate the superiority of SetRank compared with various state-of-the-art baselines.

Recent advancements in lignin valorization and biomedical applications: A patent review

2021 ◽  
Vol 15 ◽  
Author(s):  
Vandana Prasad ◽  
Lubna Siddiqui ◽  
Pawan Kumar Mishra ◽  
Adam Ekielski ◽  
Sushama Talegaonkar
Keyword(s):  
Biomedical Applications ◽  
Drug Carrier ◽  
Industrial Applications ◽  
Future Research ◽  
Lignin Valorization ◽  
New Methods ◽  
Cleaning Agents ◽  
Patent Review ◽  
Commercial Applications ◽  
Made In

Background: Synthetic polymers present disadvantages such as high cost, limited availability, safety concerns, environmental hazards and overtime accumulation in body. Lignin, an aromatic biopolymer, is highly abundant and offers various advantages including cost effectiveness, biocompatibility and biodegradability. It also possesses various pharmacological activities including antioxidant, antibacterial, anticancer and UV protection, thus lignin has become a popular biopolymer in recent years and is no more considered as bio-waste rather an extensive research is been carried out on developing it as drug carrier. Lignin also has non-biomedical applications including dispersing agents, surfactants, detergent/cleaning agents, energy storage, etc. Methods: This review compiles patents granted on production of technical lignin, different lignin therapeutic carriers and its biomedical and non-biomedical applications. The literature is collected from recent years including both articles as well as patents and is carefully analyzed and compiled in an easy to comprehend pattern for guiding future research. Results: The reviewed patents and articles highlighted the advancement made in lignin isolation and valorization. Numerous lignin nanoformulations as drug delivery agents or as standalone entities with various pharmacological actions like antibacterial, antioxidant or UV protectant have been reported. As well as industrial applications of lignin as adhesives, insulators or supercapacitors have also made lignin a biopolymer of choice. Conclusion: Lignin being a bio-inspired polymer has huge potential in commercial applications. New methods of lignin isolation from lignocellulosic biomass including physical pretreatments, solvent fraction, and chemical and biological pretreatment have been widely patented. Several micro/nano lignin formulations with improved and controllable reactivity like nanocontainers, nanocapsules, nanoparticles have also been reported recently. Also various pharmacological properties of lignin have also been explored, thus valorization of lignin is a hot topic of hour.

Sign in / Sign up

Export Citation Format

Share Document