Results 1 - 10
of
13
Learning preconditions for planning from plan traces and HTN structure
- Computational Intelligence
, 2005
"... Agreat challenge in developing planning systems for practical applications is the difficulty of acquiring the domain information needed to guide such systems. This paper describes a way to learn some of that knowledge. More specifically, the following points are discussed. (1) We introduce a theoret ..."
Abstract
-
Cited by 14 (3 self)
- Add to MetaCart
(Show Context)
Agreat challenge in developing planning systems for practical applications is the difficulty of acquiring the domain information needed to guide such systems. This paper describes a way to learn some of that knowledge. More specifically, the following points are discussed. (1) We introduce a theoretical basis for formally defining algorithms that learn preconditions for Hierarchical Task Network (HTN) methods. (2) We describe Candidate Elimination Method Learner (CaMeL), a supervised, eager, and incremental learning process for preconditions of HTN methods. We state and prove theorems about CaMeL’s soundness, completeness, and convergence properties. (3) We present empirical results about CaMeL’s convergence under various conditions. Among other things, CaMeL converges the fastest on the preconditions of the HTN methods that are needed the most often. Thus CaMeL’s output can be useful even before it has fully converged.
ECEASST Decentralized Probabilistic World Modeling with Cooperative Sensing
"... Abstract: Drawing on the projected increase in computing power, solid-state storage and network communication capacity to be available on personal mobile devices, we propose to build and maintain without prior knowledge a fully distributed decentralized large-scale model of the physical world around ..."
Abstract
-
Cited by 2 (2 self)
- Add to MetaCart
(Show Context)
Abstract: Drawing on the projected increase in computing power, solid-state storage and network communication capacity to be available on personal mobile devices, we propose to build and maintain without prior knowledge a fully distributed decentralized large-scale model of the physical world around us using probabilistic methods. We envisage that, by using the multimodal sensing capabilities of modern personal devices, such a probabilistic world model can be constructed as a collaborative effort of a community of participants, where the model data is redundantly stored on individual devices and updated and refined through short-range wireless peer-to-peer communication. Every device holds model data describing its current surroundings, and obtains model data from others when moving into unknown territory. The model represents common spatio-temporal patterns as observed by multiple participants, so that rogue participants can not easily insert false data and only patterns of general applicability dominate. This paper aims to describe – at a conceptual level – an approach for building such a distributed world model. As one possible world modeling approach, it discusses compositional hierarchies, to fuse the data from multiple sensors available on mobile devices in a bottom-up way. Furthermore, it focuses on the intertwining between building a decentralized cooperative world model and the opportunistic communication between participants.
Representations for a Complex World: Combining Distributed and Localist Representations for Learning and Planning
"... Abstract. To have agents autonomously model a complex environment, it is desirable to use distributed representations that lend themselves to neural learning. Yet developing and executing plans acting on the environment calls for abstract, localist representations of events, objects and categories. ..."
Abstract
-
Cited by 1 (0 self)
- Add to MetaCart
(Show Context)
Abstract. To have agents autonomously model a complex environment, it is desirable to use distributed representations that lend themselves to neural learning. Yet developing and executing plans acting on the environment calls for abstract, localist representations of events, objects and categories. To combine these requirements, a formalism that can express neural networks, action sequences and symbolic abstractions with the same means may be considered advantageous. We are currently exploring the use of compositional hierarchies that we treat both as Knowledge Based Artificial Neural Networks and as localist representations for plans and control structures. These hierarchies are implemented using MicroPsi node nets and used in the control of agents situated in a complex simulated environment. 1.
Unsupervised incremental learning and prediction of audio signals
- in Proceedings of 20th International Symposium on Music Acoustics
, 2010
"... The artful play with the listener’s expectations is one of the supreme skills of a gifted musician. We present a system that analyzes an audio signal in an unsupervised manner in order to generate a musical representation of it on-the-fly. The system performs the task of next note prediction using t ..."
Abstract
-
Cited by 1 (1 self)
- Add to MetaCart
(Show Context)
The artful play with the listener’s expectations is one of the supreme skills of a gifted musician. We present a system that analyzes an audio signal in an unsupervised manner in order to generate a musical representation of it on-the-fly. The system performs the task of next note prediction using the emerged representation. The main difference between our system and other existing music prediction systems is the fact that it dynamically creates the necessary representations as needed. Therefore it can adapt itself to any type of sounds, with as many timbre classes as there may be. The system consists of a conceptual clustering algorithm coupled with a modified hierarchical N-gram. The main flow of the system can be summarized in the following processing steps: 1) segmentation by transient detection, 2) timbre representation of each segment by Mel-cepstrum coefficients, 3) discretization by conceptual clustering, yielding a number of different sound classes (e.g. instruments) that can incrementally grow or shrink depending on the context resulting in a discrete sequence of sound events, 4) extraction of statistical regularities using hierarchical N-grams (Pfleger 2002), 5) prediction of continuation, and 6) sonification. The system is tested on voice recordings. We assess the robustness of the performance with respect to complexity and noise of the signal. Given that the number of estimated timbre classes is not necessarily the same as in the ground truth, we propose a performance measure (F-recall) based on pairwise matching. Finally, we sonify the predicted sequence in order to evaluate the system from a qualitative point of view. We evaluate separately the
Scaling Up Grounded Representations Hierarchically
"... Abstract We have been studying the learning of compositional hierarchies in predictive models, an area we feel is significantly underrepresented in machine learning. The aim in learning such models is to scale up automatically from fine-grained to coarser representations by identifying frequently o ..."
Abstract
- Add to MetaCart
(Show Context)
Abstract We have been studying the learning of compositional hierarchies in predictive models, an area we feel is significantly underrepresented in machine learning. The aim in learning such models is to scale up automatically from fine-grained to coarser representations by identifying frequently occurring repeated patterns, while retaining the ability to make predictions based on the statistica] regularities exhibited by these patterns. Our hierarchical learning begins with data
Machine Listening for Context-Aware Computing
, 2006
"... Machine listening is an area of study which is rapidly increasing in importance. The prolif-eration of massive sensory corpora, together with the perceptual needs of smart computa-tional devices and smart spaces has lead to this increase. Machine listening provides both a computationally cheap alter ..."
Abstract
- Add to MetaCart
(Show Context)
Machine listening is an area of study which is rapidly increasing in importance. The prolif-eration of massive sensory corpora, together with the perceptual needs of smart computa-tional devices and smart spaces has lead to this increase. Machine listening provides both a computationally cheap alternative to machine vision, and a source of information that is complementary to visual information; hence, perceptual systems which lack the abil-ity to process auditory information will in general perform less well than those which can process auditory information. Machine listening is also interesting in its own right, as re-search into computational auditory processing can help to shed light on general principles of perception, and on how our own perceptual systems work. This thesis describes machine listening research designed to solve real-world problems in perceptual and context-aware computing. This thesis makes two claims. First, it claims that machine listening technologies are well-suited to the task of providing context awareness in real-world computational systems, whether these systems are intended to provide operational cues to smart devices or spaces,
Partial N-grams
"... uirement) partial n-grams can provide better performance. We performed experiments using book1 from the Calgary compression corpus [1], a Thomas Hardy novel, transformed to a 26-letter alphabet (using letters as the basic symbols, not words). We examined standard predict-the-next-symbol inference, ..."
Abstract
- Add to MetaCart
(Show Context)
uirement) partial n-grams can provide better performance. We performed experiments using book1 from the Calgary compression corpus [1], a Thomas Hardy novel, transformed to a 26-letter alphabet (using letters as the basic symbols, not words). We examined standard predict-the-next-symbol inference, though in general we are interested in arbitrary prediction patterns, such as predicting a middle symbol from context on both sides or simultaneously predicting multiple symbols [5]. (This generality necessitates representing estimates for the full joint distribution rather than the conditional distribution.) Accuracy and entropy were measured using a held out test set consisting of the last 10,000 characters. Accuracy is the standard predictive accuracy common in machine learning, the proportion of times the correct symbol was predicted by the model. Entropy here means the standard measure referred to as the entropy of the test data given the model, or the cross-entropy. We expected partial
On-Line Learning of Undirected Sparse n-grams
"... n-grams are simple learning models considered state-of-the-art in many sequential domains. They suffer from an exponential number of parameters in their width, n. We introduce undirected sparse n-grams, which store probability estimates only for some n-tuples from the unconditional joint distrib ..."
Abstract
- Add to MetaCart
n-grams are simple learning models considered state-of-the-art in many sequential domains. They suffer from an exponential number of parameters in their width, n. We introduce undirected sparse n-grams, which store probability estimates only for some n-tuples from the unconditional joint distribution, specifically the most frequent. Experimental results show this sparse version can outperform a narrower n-gram with the same number of parameters. In addition, our models predict equally well forward, backward, or in any combination. We present an on-line learning algorithm which induces both model parameters and structure (which patterns to include), despite a priori uncertainty about which patterns are frequent. Dynamic pattern inclusion complicates probability estimation during on-line learning, but we present and examine several solutions, including a Bayesian approach. Lastly, we describe multiwidth combinations of sparse n-grams that solve two important problems with the simple sparse models, and which are useful for hierarchical identification and composition of repeated substructure in data. These hierarchical sparse n-grams can be learned on-line with no fixed bound on width, using less memory than that taken by the training data. 1
PLANNING BY EXAMINING DECOMPOSITIONAL PLAN TRACES
, 2006
"... Knowledge-based, hand-tailorable planning systems are the most promising planning systems to solve real-wold planning problems. A great challenge in using any such knowledge-based planning system in real world is the difficulty of acquiring the domain knowledge needed to guide such a system. The obj ..."
Abstract
- Add to MetaCart
(Show Context)
Knowledge-based, hand-tailorable planning systems are the most promising planning systems to solve real-wold planning problems. A great challenge in using any such knowledge-based planning system in real world is the difficulty of acquiring the domain knowledge needed to guide such a system. The objective of this disser-tation is to investigate ways to acquire parts or all of this information automatically in the context of Hierarchical Task Network (HTN) planning. Knowledge acquisi-tion is done by examining the decisions made by an expert problem solver while solving planning problems in a given HTN domain. This dissertation first describes a general framework to define what an HTN domain learner is, what its inputs and outputs are, how to evaluate such a learner, and what soundness, completeness and convergence mean in this context. Afterwards, two different HTN domain learning algorithms CaMeL and HDL are discussed. These two algorithms are then extended to handle noise in training samples, and to help HTN planners to start planning before full convergence is achieved by CaMeL or HDL. These extensions result in a family of different HTN domain learning algorithms, each of which can be useful un-
Multimodal Estimation of User Interruptibility for Smart Mobile Telephones ABSTRACT
"... Context-aware computer systems are characterized by the ability to consider user state information in their decision logic. One example application of context-aware computing is the smart mobile telephone. Ideally, a smart mobile telephone should be able to consider both social factors (i.e. known r ..."
Abstract
- Add to MetaCart
(Show Context)
Context-aware computer systems are characterized by the ability to consider user state information in their decision logic. One example application of context-aware computing is the smart mobile telephone. Ideally, a smart mobile telephone should be able to consider both social factors (i.e. known relationships between contactor and contactee) and environmental factors (i.e. the contactee’s current locale and activity) when deciding how to handle an incoming request for communication. Toward providing this kind of user state information and improving the ability of the mobile phone to handle calls intelligently, we present work on inferring environmental factors from sensory data and using this information to predict user interruptibility. Specifically, we learn the structure and parameters of a user state model from continuous ambient audio and visual information from periodic still images, and attempt to associate the learned states with user-reported interruptibility levels. We report experimental results using this technique on real data, and show how such an approach can allow for adaptation to specific user preferences.
