Filters








169,155 Hits in 4.5 sec

Quantifying Relevance in Learning and Inference [article]

Matteo Marsili, Yasser Roudi
2022
The relevance, as we define it here, quantifies the amount of information that a dataset or the internal representation of a learning machine contains on the generative model of the data.  ...  These applications push statistical inference into uncharted territories where data is high-dimensional and scarce, and prior information on "true" models is scant if not totally absent.  ...  Acknowledgments We're greatly indebted to our collaborators, Ryan Cubero, Odilon Duranthon, Ariel Haimovici, Silvio Franz, Silvia Grigolon, Jungyo Jo, Iacopo Mastromatteo, Juyong Song and Rongrong (Emily  ... 
doi:10.48550/arxiv.2202.00339 fatcat:fbrz7vtwifbn7gp4j4q3jrr2bu

Pragmatic, linguistic and cognitive factors in young children's development of quantity, relevance and word learning inferences

Elspeth WILSON, Napoleon KATSOS
2021 Journal of Child Language  
We investigated three- to five-year-old English-speaking children's (N = 71) performance in ad hoc quantity, scalar quantity and relevance implicatures, as well as word learning by exclusion inferences  ...  Children's pragmatic abilities improved with age, with word learning by exclusion acquired first, followed by relevance and ad hoc quantity implicatures, and finally scalar quantity implicatures.  ...  We are grateful to the families and schools who took part in the study, and Becky Brooks for assistance with data entry.  ... 
doi:10.1017/s0305000921000453 fatcat:ripfqmu565atvho7ryro4gfm5e

Choice variability and suboptimality in uncertain environments

Valentin Wyart, Etienne Koechlin
2016 Current Opinion in Behavioral Sciences  
biases in inference per se.  ...  This variability is usually hypothesized as noise at the periphery of inferential processes, namely sensory noise in perceptual tasks and stochastic exploration in reward-guided learning, or as suboptimal  ...  Acknowledgements References and recommended reading  ... 
doi:10.1016/j.cobeha.2016.07.003 fatcat:2f2vizehizce5dy6fgdupo5o4i

Why the Decision Theoretic Perspective Misrepresents Frequentist Inference: 'Nuts and Bolts' vs. Learning from Data [article]

Aris Spanos
2016 arXiv   pre-print
It provides the Bayesian approach with a theory of optimal inference, but it misrepresents the theory of optimal frequentist inference by framing inferences solely in terms of the universal quantifier  ...  'for all values of theta in the parameter space'.  ...  At first sight the quantifier ∀θ∈Θ seems rather innocuous and natural in the context of statistical inference.  ... 
arXiv:1211.0638v3 fatcat:7z6463vizbf4jdyapxhjsnjvyi

Graph Traversal Methods for Reasoning in Large Knowledge-Based Systems

Abhishek Sharma, Kenneth Forbus
2013 PROCEEDINGS OF THE THIRTIETH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE AND THE TWENTY-EIGHTH INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE CONFERENCE  
In this paper, we discuss two ways in which heuristic graph traversal methods can be used to generate plausible inference chains.  ...  Finally, we demonstrate through experiments that these methods lead to significant improvement in accuracy for both Q/A and script construction.  ...  Five sets of questions were selected based on the availability of ground facts in KB and their relevance in learning by reading [Forbus et al 2007] .  ... 
doi:10.1609/aaai.v27i1.8473 fatcat:ogmd75q6iba23afsdfcftucjha

Accessing the unsaid: The role of scalar alternatives in children's pragmatic inference

David Barner, Neon Brooks, Alan Bale
2011 Cognition  
Four-year-olds were shown pictures in which two objects fit a description and a third object did not, and were asked to judge the truth value of statements that relied on context-independent alternatives  ...  scalar implicature result from a lack of knowledge of the relevant alternatives.  ...  In contrast, normal children never learn to recite a sequence of quantifiers like some, many, most, all, etc.  ... 
doi:10.1016/j.cognition.2010.10.010 pmid:21074147 fatcat:ckn7is22gvblldmqilfponmoge

Teaching and Learning Guide for: Pragmatics: From Theory to Experiment and Back Again

Napoleon Katsos, Chris Cummins
2010 Language and Linguistics Compass  
They interpret this as evidence that the Relevance Theory account (in which the derivation of inferences is contextual and effortful) is favoured over the Default Inference account (in which the derivation  ...  This paper tests the competing predictions of Default Inference and Relevance Theory accounts of scalar implicature, using a sentence verification paradigm and underinformative statements.  ... 
doi:10.1111/j.1749-818x.2010.00247.x fatcat:y36oaernqbfkblbavhd73aqtia

Towards unobtrusive Parkinson's disease detection via motor symptoms severity inference from multimodal smartphone-sensor data

Dimitrios Iakovakis, Stelios Hadjidimitriou, Vasileios Charisis, Konstantinos Kyritsis, Alexandros Papadopoulos, Michael Stadtschnitzer, Hagen Jaeger, Ioannis Dagklis, Sevasti Bostantjopoulou, Zoe Katsarou, Lisa Klingelhoefer, Simone Mayer (+7 others)
2019 Zenodo  
iPrognosis Android smartphone application, for relevant motor symptoms severity inference.  ...  Objective: To provide clinically-corroborated evidence of the Parkinson's disease (PD) diagnostic potential of machine learning-based approaches for motor symptoms severity inference via multimodal data  ...  iPrognosis Android smartphone application, for relevant motor symptoms severity inference.  ... 
doi:10.5281/zenodo.3675351 fatcat:owyph3ptongwfo5owy5jmumtni

Machine learning-assisted identification of bioindicators predicts medium-chain carboxylate production performance of an anaerobic mixed culture

Bin Liu
2020 International Chain Elongation Conference 2020  
An abstract can be found in the right column.  ...  The inferred bioindicators may delineate their relevance to the 55 enhanced C6/C8 productivity in the chain elongation process, manipulated 56 by HRT decline.  ...  Since C6 and C8 were the 29 target products, it was relevant to manipulate the reactor microbiome for in both reactors at the same HRT (P > 0.05).  ... 
doi:10.18174/icec2020.18013 fatcat:rib3tgqdgvhgxn6qbc4z6nfdd4

Why the Decision‐Theoretic Perspective Misrepresents Frequentist Inference: Revisiting Stein's Paradox and Admissibility [chapter]

Aris Spanos
2017 Advances in Statistical Methodologies and Their Application to Real Problems  
In contrast, the theory of optimal frequentist inference is framed entirely in terms of the capacity of the procedure to pinpoint θ * : The inappropriateness of the quantifier ∀θ ∈ Θ calls into question  ...  Decision-theoretic and Bayesian rules are considered optimal when they minimize the expected loss "for all possible values of θ in Θ" ½∀θ ∈ Θ; irrespective of what the true value θ * [state of Nature]  ...  Frequentist inference and learning from data The objectives and underlying reasoning of frequentist inference are inadequately discussed in the statistics literature.  ... 
doi:10.5772/65720 fatcat:wtbtq2ukp5aodptf4q6bras7li

Autonomous science platforms and question-asking machines

Kevin H. Knuth, Julian L. Center
2010 2010 2nd International Workshop on Cognitive Information Processing  
In this paper we view these platforms as question-asking machines and introduce a paradigm based on the scientific method, which couples the processes of inference and inquiry to form a model-based learning  ...  In simple cases, the relevance is proportional to the entropy. This data is then analyzed by the inference engine, which updates the state of knowledge of the instrument.  ...  In addition, K.H.  ... 
doi:10.1109/cip.2010.5604217 dblp:conf/cogip/KnuthC10 fatcat:dmavuba4qndepkg27otnl3ne7q

An Information-Theoretic Analysis of the Cost of Decentralization for Learning and Inference under Privacy Constraints

Sharu Theresa Jose, Osvaldo Simeone
2022 Entropy  
The cost of decentralization for learning and/or inference is shown to be quantified in terms of conditional mutual information terms involving features and label variables.  ...  A fundamental theoretical question in this setting is how to quantify the cost, or performance loss, of decentralization for learning and/or inference.  ...  Our main goal is to quantify, using information-theoretic metrics, the benefits of cooperation for learning and/or inference.  ... 
doi:10.3390/e24040485 pmid:35455148 pmcid:PMC9030603 fatcat:scby5gqo3nbojlf76joqavt5ee

Scoring Qualitative Informal Learning Dialogue: The SQuILD Method for Measuring Museum Learning Talk

Jessica Roberts, Leilah Lyons
2017 International Conference on Computer Supported Collaborative Learning  
In order to evaluate learning outcomes for CSCL exhibits, we present a method for quantifying idiosyncratic social learning, Scoring Qualitative Informal Learning Dialogue (SQuILD), and demonstrate how  ...  Museums are increasingly developing computer-supported collaborative learning experiences and are in need of methods for evaluating the educational value of such exhibits.  ...  These relevance categories are assigned numerical weights in order to quantify and compare the substance of visitor talk across conditions.  ... 
dblp:conf/cscl/RobertsL17 fatcat:dnauynbyuzbthlvwtdnyr7lggm

On the importance of being critical

Matteo Marsili
2020 EurophysicsNews  
Being critical, i.e. able to process and distill relevant information, is crucial for living systems. Learning distinguishes living from inanimate matter.  ...  Quantifying this distinction may provide a "life meter" [1] that, for example, can allow us to detect alien life forms in astrobiology.  ...  He's interested in understanding collective phenomena in different disciplines (physics, biology, economics and finance, statistical inference and learning) with statistical physics methods. .  ... 
doi:10.1051/epn/2020508 fatcat:vahvvs3oenbxrca7qk5bfkevry

Scientific Inference With Interpretable Machine Learning: Analyzing Models to Learn About Real-World Phenomena [article]

Timo Freiesleben, Gunnar König, Christoph Molnar, Alvaro Tejero-Cantero
2022 arXiv   pre-print
Interpretable machine learning (IML) is concerned with the behavior and the properties of machine learning models.  ...  Our phenomenon-centric approach to IML in science clarifies: the opportunities and limitations of IML for inference; that conditional not marginal sampling is required; and, the conditions under which  ...  described in the paper.  ... 
arXiv:2206.05487v1 fatcat:cc3ws2jvfjg6vkniftzcrfqy6q
« Previous Showing results 1 — 15 out of 169,155 results