scholarly journals Hiding Satisfying Assignments: Two are Better than One

2005 ◽  
Vol 24 ◽  
pp. 623-639 ◽  
Author(s):  
D. Achlioptas ◽  
H. Jia ◽  
C. Moore
Keyword(s):  
Experimental Evidence ◽  
Truth Assignment ◽  
Satisfiability Solvers ◽  
Basic Model ◽  
Better Than

The evaluation of incomplete satisfiability solvers depends critically on the availability of hard satisfiable instances. A plausible source of such instances consists of random k-SAT formulas whose clauses are chosen uniformly from among all clauses satisfying some randomly chosen truth assignment A. Unfortunately, instances generated in this manner tend to be relatively easy and can be solved efficiently by practical heuristics. Roughly speaking, for a number of different algorithms, A acts as a stronger and stronger attractor as the formula's density increases. Motivated by recent results on the geometry of the space of satisfying truth assignments of random k-SAT and NAE-k-SAT formulas, we introduce a simple twist on this basic model, which appears to dramatically increase its hardness. Namely, in addition to forbidding the clauses violated by the hidden assignment A, we also forbid the clauses violated by its complement, so that both A and compliment of A are satisfying. It appears that under this "symmetrization" the effects of the two attractors largely cancel out, making it much harder for algorithms to find any truth assignment. We give theoretical and experimental evidence supporting this assertion.

Practical Implications of Modeling Commercial Vehicle Empty Trips

2003 ◽  
Vol 1833 (1) ◽  
pp. 87-94 ◽  
Author(s):  
José Holguin-Veras ◽  
Ellen Thorson
Keyword(s):  
Parameter Estimation ◽  
Commercial Vehicle ◽  
Complex Case ◽  
Statistical Distributions ◽  
Estimation Errors ◽  
Theoretical Derivation ◽  
Basic Model ◽  
Order Of Magnitude ◽  
Practical Implications ◽  
Better Than

Implications of modeling commercial vehicle empty trips are discussed, a theoretical derivation for parameter estimation is provided, and insight is given into the order of magnitude of estimation errors because of the improper modeling of commercial vehicle empty trips. A set of relatively simple cases was designed to illustrate the most important implications. Also addressed are estimation errors from using naïve approaches to compensate for the lack of explicit modeling of empty trips and the errors associated with more advanced empty trip models. In the simplest simulation, directional errors for a basic complementary model were from three to six times fewer than those for the naïve models. In the more complex case, a more sophisticated complementary model performed slightly better than the basic model and both complementary models were considerably better than the naïve approaches. The directional errors for the naïve models were four to seven times greater than those for the complementary models. Moreover, an analysis of the statistical distributions of the errors indicated that the complementary models had higher probabilities of producing accurate results, whereas the naïve approaches had higher probabilities of producing very large errors. These analyses indicate that the naïve approaches translate into significant errors in directional-traffic estimates. For that reason, their use should be discontinued in favor of the more advanced models presented.

scholarly journals Local Regularizer Improves Generalization

2020 ◽  
Vol 34 (04) ◽  
pp. 6861-6868 ◽  
Author(s):  
Yikai Zhang ◽  
Hui Qu ◽  
Dimitris Metaxas ◽  
Chao Chen
Keyword(s):  
Deep Learning ◽  
Theoretical Analysis ◽  
Experimental Evidence ◽  
Gradient Descent ◽  
Stochastic Gradient ◽  
Stochastic Gradient Descent ◽  
Training Algorithms ◽  
Training Methods ◽  
Theoretical Understanding ◽  
Better Than

Regularization plays an important role in generalization of deep learning. In this paper, we study the generalization power of an unbiased regularizor for training algorithms in deep learning. We focus on training methods called Locally Regularized Stochastic Gradient Descent (LRSGD). An LRSGD leverages a proximal type penalty in gradient descent steps to regularize SGD in training. We show that by carefully choosing relevant parameters, LRSGD generalizes better than SGD. Our thorough theoretical analysis is supported by experimental evidence. It advances our theoretical understanding of deep learning and provides new perspectives on designing training algorithms. The code is available at https://github.com/huiqu18/LRSGD.

scholarly journals Prevention Is Better Than Cure: Experimental Evidence From Milk Fever Incidence in Dairy Animals of Haryana, India

2021 ◽  
Author(s):  
A G Adeeth Cariappa ◽  
B S Chandel ◽  
Gopal Sankhala ◽  
Veena Mani ◽  
Sendhil R ◽  
...  
Keyword(s):  
Experimental Evidence ◽  
Milk Fever ◽  
Dairy Animals ◽  
Better Than

scholarly journals Supervised attention for answer selection in community question answering

2020 ◽  
Vol 9 (2) ◽  
pp. 203
Author(s):  
Thanh Thi Ha ◽  
Atsuhiro Takasu ◽  
Thanh Chinh Nguyen ◽  
Kiem Hieu Nguyen ◽  
Van Nha Nguyen ◽  
...  
Keyword(s):  
Language Processing ◽  
Question Answering ◽  
Irrelevant Information ◽  
Social Question ◽  
Community Question Answering ◽  
Basic Model ◽  
Proposed Model ◽  
Questions And Answers ◽  
Word Attention ◽  
Better Than

<span class="fontstyle0">Answer selection is an important task in Community Question Answering (CQA). In recent years, attention-based neural networks have been extensively studied in various natural language processing problems, including question answering. This paper explores </span><span class="fontstyle2">matchLSTM </span><span class="fontstyle0">for answer selection in CQA. A lexical gap in CQA is more challenging as questions and answers typical contain multiple sentences, irrelevant information, and noisy expressions. In our investigation, word-by-word attention in the original model does not work well on social question-answer pairs. We propose integrating supervised attention into </span><span class="fontstyle2">matchLSTM</span><span class="fontstyle0">. Specifically, we leverage lexical-semantic from external to guide the learning of attention weights for question-answer pairs. The proposed model learns more meaningful attention that allows performing better than the basic model. Our performance is among the top on SemEval datasets.</span> <br /><br />

scholarly journals Graph biased feature selection of genes is better than random for many genes

2020 ◽  
Author(s):  
Jake Crawford ◽  
Casey S. Greene
Keyword(s):  
Gene Expression ◽  
Feature Selection ◽  
Experimental Evidence ◽  
Recent Work ◽  
Random Model ◽  
Interaction Networks ◽  
Random Networks ◽  
Low Degree ◽  
Selection Of ◽  
Better Than

AbstractRecent work suggests that gene expression dependencies can be predicted almost as well by using random networks as by using experimentally derived interaction networks. We hypothesize that this effect is highly variable across genes, as useful and robust experimental evidence exists for some genes but not others. To explore this variation, we take the k-core decomposition of the STRING network, and compare it to a degree-matched random model. We show that when low-degree nodes are removed, expression dependencies in the remaining genes can be predicted better by the resulting network than by the random model.

The Impact of Psychological Traits on Performance Under Pressure—Experimental Evidence of Handball Penalties

2021 ◽  
pp. 152700252110271
Author(s):  
Christoph Bühren ◽  
Lisa Träger
Keyword(s):  
Field Experiment ◽  
Experimental Evidence ◽  
Psychological Traits ◽  
Sequential Games ◽  
First Mover Advantage ◽  
The Impact ◽  
First Mover ◽  
Better Than ◽  
Performance Under Pressure

Our field experiment analyzes the influence of psychological traits on performance in sequential games. It uses handball penalties thrown under individual, team, or tournament incentives in the ABBA sequence. Considering the single moves of these games, player A and player B are taking turns in being the first-mover. We find no significant first-mover advantage. However, we observe that player A performs better than player B under tournament incentives and if he or she is confident enough.

Left is Better Than Right for Reducing Nondeterminism of NFAs

2021 ◽  
pp. 1-20
Author(s):  
Sang-Ki Ko ◽  
Yo-Sub Han
Keyword(s):  
Experimental Evidence ◽  
Equivalence Class ◽  
Polynomial Time ◽  
Minimization Problem ◽  
The Right ◽  
Present Experimental Evidence ◽  
Better Than

We study the NFA reductions by invariant equivalences and preorders. It is well-known that the NFA minimization problem is PSPACE-complete. Therefore, there have been many approaches to reduce the size of NFAs in low polynomial time by computing invariant equivalence or preorder relation and merging the states within same equivalence class. Here we consider the nondeterminism reduction of NFAs by invariant equivalences and preorders. We, in particular, show that computing equivalence and preorder relation from the left is more useful than the right for reducing the degree of nondeterminism in NFAs. We also present experimental evidence for showing that NFA reduction from the left achieves the better reduction of nondeterminism than reduction from the right.

Altruists with Green Beards

2005 ◽  
Vol 27 (1) ◽  
Author(s):  
Ernst Fehr ◽  
Urs Fischbacher
Keyword(s):  
Experimental Evidence ◽  
Evolution Of Cooperation ◽  
Conditional Cooperation ◽  
Face To Face ◽  
Prisoners Dilemma ◽  
Phenotypic Features ◽  
Better Than

AbstractIf cooperative dispositions are associated with unique phenotypic features (’green beards’), cooperative individuals can be identified. Therefore, cooperative individuals can avoid exploitation by defectors by cooperating exclusively with other cooperative individuals; consequently, cooperators flourish and defectors die out. Experimental evidence suggests that subjects, who are given the opportunity to make promises in face-to-face interactions, are indeed able to predict the partner’s behavior better than chance in a subsequent Prisoners’ Dilemma. This evidence has been interpreted as evidence in favor of green beard approaches to the evolution of human cooperation. Here we argue, however, that the evidence does not support this interpretation. We show, in particular, that the existence of conditional cooperation renders subjects' choices in the Prisoners’ Dilemma predictable. However, although subjects predict behavior better than chance, selfish individuals earn higher incomes than conditional cooperators. Thus, although subjects may predict other players’ choices better than chance evolution favors the selfish subjects, i.e., the experimental evidence does not support the green beard approach towards the evolution of cooperation.

Vibration-Based Outlier Detection on High Dimensional Data

2016 ◽  
Vol 25 (03) ◽  
pp. 1650013
Author(s):  
Shuyin Xia ◽  
Guoyin Wang ◽  
Hong Yu ◽  
Qun Liu ◽  
Jin Wang
Keyword(s):  
Outlier Detection ◽  
Time Complexity ◽  
State Of The Art ◽  
High Dimensional Data ◽  
Difficult Problem ◽  
Index Structure ◽  
High Dimensional ◽  
Basic Model ◽  
Traditional Approaches ◽  
Better Than

Outlier detection is a difficult problem due to its time complexity being quadratic or cube in most cases, which makes it necessary to develop corresponding acceleration algorithms. Since the index structure (c.f. R tree) is used in the main acceleration algorithms, those approaches deteriorate when the dimensionality increases. In this paper, an approach named VBOD (vibration-based outlier detection) is proposed, in which the main variants assess the vibration. Since the basic model and approximation algorithm FASTVBOD do not need to compute the index structure, their performances are less sensitive to increasing dimensions than traditional approaches. The basic model of this approach has only quadratic time complexity. Furthermore, accelerated algorithms decrease time complexity to [Formula: see text]. The fact that this approach does not rely on any parameter selection is another advantage. FASTVBOD was compared with other state-of-the-art algorithms, and it performed much better than other methods especially on high dimensional data.

Sign in / Sign up

Export Citation Format

Share Document