检索结果-内蒙古大学图书馆

23rd International Conference on Artificial Intelligence in Education (AIED)

作者： Tang, Leonard Ke, Elizabeth Singh, Nikhil Feng, Bo Austin, Derek Verma, Nakul Drori, Iddo MIT 77 Massachusetts Ave Cambridge MA 02139 USA Harvard Univ Cambridge MA 02138 USA Columbia Univ New York NY 10025 USA

ISBN: (纸本)9783031116476;9783031116469

We use probabilistic program synthesis to solve questions in MIT and Harvard Probability and Statistics courses. Traditional approaches using the latest GPT-3 language model without program synthesis achieve a solve rate of 0.2 in these classes. In contrast, by turning course questions into probabilistic programs using the latest program synthesis Transformer, OpenAI Codex, and executing the programs, our solve rates are 0.9 and 0.88, which are on par with human performance.

关键词： Transformers Program synthesis probabilistic programming Few-shot learning

来源：评论

学校读者我要写书评

暂无评论

Compiling Markov Chain Monte Carlo Algorithms for probabilistic Modeling 2017

Compiling Markov Chain Monte Carlo Algorithms for Probabilis...

引用

38th ACM SIGPLAN Conference on programming Language Design and Implementation (PLDI)

作者： Huang, Daniel Tristan, Jean-Baptiste Morrisett, Greg Harvard Univ Cambridge MA 02138 USA Oracle Labs Burlington MA USA Cornell Univ Ithaca NY USA

ISBN: (纸本)9781450349888

The problem of probabilistic modeling and inference, at a high-level, can be viewed as constructing a (model, query, inference) tuple, where an inference algorithm implements a query on a model. Notably, the derivation of inference algorithms can be a difficult and error-prone task. Hence, researchers have explored how ideas from probabilistic prog ramming can be applied. In the context of constructing these tuples, probabilistic programming can be seen as taking a language-based approach to probabilistic modeling and inference. For instance, by using (1) appropriate languages for expressing models and queries and (2) devising inference techniques that operate on encodings of models (and queries) as program expressions, the task of inference can be automated. In this paper, we describe a compiler that transforms a probabilistic model written in a restricted modeling language and a query for posterior samples given observed data into a Markov Chain Monte Carlo (MCMC) inference algorithm that implements the query. The compiler uses a sequence of intermediate languages (ILs) that guide it in gradually and successively refining a declarative specification of a probabilistic model and the query into an executable MCMC inference algorithm. The compilation strategy produces composable MCMC algorithms for execution on a CPU or GPU.

关键词： probabilistic programming intermediate languages Markov-chain Monte Carlo kernels

来源：评论

学校读者我要写书评

暂无评论

Contextual Equivalence for a probabilistic Language with Continuous Random Variables and Recursion

引用

PROCEEDINGS OF THE ACM ON programming LANGUAGES-PACMPL 2018年第ICFP期2卷 1–30页

作者： Wand, Mitchell Culpepper, Ryan Giannakopoulos, Theophilos Cobb, Andrew Northeastern Univ Coll Comp & Informat Sci 360 Huntington AveRoom 202WVH Boston MA 02115 USA Czech Tech Univ Fac Informat Technol Thakurova 9 Prague 16000 Czech Republic BAE Syst FAST Labs 600 Dist Ave Burlington MA 01803 USA

We present a complete reasoning principle for contextual equivalence in an untyped probabilistic language. The language includes continuous (real-valued) random variables, conditionals, and scoring. It also includes recursion, since the standard call-by-value fixpoint combinator is expressible. We demonstrate the usability of our characterization by proving several equivalence schemas, including familiar facts from lambda calculus as well as results specific to probabilistic programming. In particular, we use it to prove that reordering the random draws in a probabilistic program preserves contextual equivalence. This allows us to show, for example, that (let x = e(1) in let y = e(2) in e(0)) =(ctx) (let y = e(2) in let x = e(1) in e(0)) (provided x does not occur free in e(2) and y does not occur free in e(1)) despite the fact that e(1) and e(2) may have sampling and scoring effects.

关键词： probabilistic programming logical relations contextual equivalence

来源：评论

学校读者我要写书评

暂无评论

Formally Verified Samplers from probabilistic Programs with Loops and Conditioning

引用

PROCEEDINGS OF THE ACM ON programming LANGUAGES-PACMPL 2023年第PLDI期7卷 1-24页

作者： Bagnall, Alexander Stewart, Gordon Banerjee, Anindya Ohio Univ Athens OH 45701 USA BedRock Syst Inc Boston MA USA IMDEA Software Inst Madrid Spain

We present Zar: a formally verified compiler pipeline from discrete probabilistic programs with unbounded loops in the conditional probabilistic guarded command language (cpGCL) to proved-correct executable samplers in the random bit model. We exploit the key idea that all discrete probability distributions can be reduced to unbiased coin-flipping schemes. The compiler pipeline first translates cpGCL programs into choice-fix trees, an intermediate representation suitable for reduction of biased probabilistic choices. Choice-fix trees are then translated to coinductive interaction trees for execution within the random bit model. The correctness of the composed translations establishes the sampling equidistribution theorem: compiled samplers are correct wrt. the conditional weakest pre-expectation semantics of cpGCL source programs. Zar is implemented and fully verified in the Coq proof assistant. We extract verified samplers to OCaml and Python and empirically validate them on a number of illustrative examples.

关键词： probabilistic programming Verified Compilers

来源：评论

学校读者我要写书评

暂无评论

Surprisal, Liking, and Musical Affect 7th

Surprisal, Liking, and Musical Affect

引用

7th Biennial International Conference for Mathematics and Computation in Music (MCM)

作者： Fram, Noah R. CCRMA 660 Lomita Dr Stanford CA 94305 USA

ISBN: (纸本)9783030213923;9783030213916

Formulation and processing of expectation has long been viewed as an essential component of the emotional, psychological, and neurological response to musical events. There are multiple theories of musical expectation, ranging from a broad association between expectation violation and musical affect to precise descriptions of neurocognitive networks that contribute to the perception of surprising stimuli. In this paper, we propose a probabilistic model of musical expectation that relies on the recursive updating of listeners' conditional predictions of future events in the musical stream. This model is defined in terms of cross-entropy, or information content given a prior model. A probabilistic program implementing some aspects of this model with melodies from Bach chorales is shown to support the hypothesized connection between the evolution of surprisal through a piece and affective arousal, indexed by the spread of possible deviations from the expected play count.

关键词： Affect Entropy Music Perception probabilistic programming Surprisal

来源：评论

学校读者我要写书评

暂无评论

PRAGMATIC APPROACHES TO OPTIMIZATION WITH RANDOM YIELD COEFFICIENTS

引用

FOREST SCIENCE 1995年第3期41卷 501-512页

作者： HOF, J BEVERS, M PICKENS, J MICHIGAN TECHNOL UNIV HOUGHTONMI 49931

This paper discusses practical methods for handling normally distributed random technical (yield) coefficients in linear programs that optimize natural resource allocation and scheduling, These methods are practical in the sense that they are applicable to large-scale real world models and do not require nonlinear solution methods. The paper begins with a description and demonstration of postoptimization approaches that are applicable to large, linear problems, and then explores methods for reducing overall risk through land allocation diversification, A central theme of the paper is the importance of providing some sort of allowance for uncertainty when presenting optimization results, which promotes a more realistic view of the problem by analysts and decision makers alike.

关键词： LINEAR programming CHANCE CONSTRAINTS probabilistic programming

来源：评论

学校读者我要写书评

暂无评论

A SURROGATE FOR LINEAR-PROGRAMS WITH RANDOM REQUIREMENTS

引用

EUROPEAN JOURNAL OF OPERATIONAL RESEARCH 1988年第3期34卷 399-402页

作者： GROVE, MA UNIV OREGON DEPT ECONEUGENEOR 97403 USA

An approach to linear programs with random requirements is suggested. The procedure involves choosing actions which minimize the expected value of a certain loss function. These actions are then taken as goals, and op... 详细信息

关键词： goal programming probabilistic programming

来源：评论

学校读者我要写书评

暂无评论

Authoring Human Simulators via probabilistic Functional Reactive Program Synthesis 22

Authoring Human Simulators via Probabilistic Functional Reac...

引用

17th Annual ACM/IEEE International Conference on Human-Robot Interaction (HRI)

作者： Chung, Michael Jae-Yoon Cakmak, Maya Univ Washington Comp Sci & Engn Seattle WA 98195 USA

ISBN: (数字)9781665407311

ISBN: (纸本)9781665407311

One of the core challenges in creating interactive behaviors for social robots is testing. Programs implementing the interactive behaviors require real humans to test and this requirement makes testing of the programs extremely expensive. To address this problem, human-robot interaction researchers in the past proposed using human simulators. However, human simulators are tedious to set up and context-dependent and therefore are not widely used in practice. We propose a program synthesis approach to building human simulators for the purpose of testing interactive robot programs. Our key ideas are (1) representing human simulators as probabilistic functional reactive programming programs and (2) using probabilistic inference for synthesizing human simulator programs. Programmers then will be able to build human simulators by providing interaction traces between a robot and a human or two humans which they can later use to test interactive robot programs and improve or tweak as needed.

关键词： probabilistic programming program synthesis end-user programming social robots

来源：评论

学校读者我要写书评

暂无评论

Understanding Human Generated Decision Data 1

引用

10th Annual International symposium on Business Modeling and Software Design (BMSD)

作者： Silvander, Johan Blekinge Inst Technol Software Engn Res Lab Sweden Karlskrona Sweden

ISBN: (数字)9783030523060

ISBN: (纸本)9783030523060;9783030523053

In order to design intent-driven systems, the understanding of how the data is generated is essential. Without the understanding of the data generation process, it is not possible to use interventions, and counterfactuals. Interventions, and counterfactuals, are useful tools in order to achieve an artificial intelligence which can improve the system itself. We will create an understanding, and a model, of how data about decisions are generated, as well as used, by human decision makers. The research data were collected with the help of focus group interviews, and questionnaires. The models were built and evaluated with the help of, bayesian statistics, probability programming, and discussions with the practitioners. When we are combining, probabilistic programming models, extended machine learning algorithms, and data science processes, into a directed acyclic graph, we can mimic the process of human generated decision data. We believe the usage of a directed acyclic graph, to combine the functions and models, is a good base for mimic human generated decision data. Our next step is to evaluate if flow-based programming can be used as a framework for realization of components, useful in intent-driven systems.

关键词： Human decisions Bayesian statistics probabilistic programming

来源：评论

学校读者我要写书评

暂无评论

Expressing and Verifying probabilistic Assertions 14

Expressing and Verifying Probabilistic Assertions

引用

35th ACM SIGPLAN Conference on programming Language Design and Implementation (PLDI)

作者： Sampson, Adrian Panchekha, Pavel Mytkowicz, Todd McKinley, Kathryn S. Grossman, Dan Ceze, Luis Univ Washington Seattle WA 98195 USA

ISBN: (纸本)9781450327848

Traditional assertions express correctness properties that must hold on every program execution. However, many applications have probabilistic outcomes and consequently their correctness properties are also probabilistic (e. g., they identify faces in images, consume sensor data, or run on unreliable hardware). Traditional assertions do not capture these correctness properties. This paper proposes that programmers express probabilistic correctness properties with probabilistic assertions and describes a new probabilistic evaluation approach to efficiently verify these assertions. probabilistic assertions are Boolean expressions that express the probability that a property will be true in a given execution rather than asserting that the property must always be true. Given either specific inputs or distributions on the input space, probabilistic evaluation verifies probabilistic assertions by first performing distribution extraction to represent the program as a Bayesian network. probabilistic evaluation then uses statistical properties to simplify this representation to efficiently compute assertion probabilities directly or with sampling. Our approach is a mix of both static and dynamic analysis: distribution extraction statically builds and optimizes the Bayesian network representation and sampling dynamically interprets this representation. We implement our approach in a tool called MAYHAP for C and C++ programs. We evaluate expressiveness, correctness, and performance of MAYHAP on programs that use sensors, perform approximate computation, and obfuscate data for privacy. Our case studies demonstrate that probabilistic assertions describe useful correctness properties and that MAYHAP efficiently verifies them.

关键词： Languages Reliability probabilistic programming approximate computing data obfuscation differential privacy sensors symbolic execution

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：