Abstract: / 26 EVALUATION IS A MAINSTREAM ACTIVITY IN HCI. For many years we saw the emergence of a plethora of techniques to measure user-orientated quality assessment of technology: usability, satisfaction, efficiency, effectiveness, learnability, usefulness, and so on. In recent times, however, the discussion seems to have moved on. Issues surrounding the wider relationship between people and technology or the user experience are popular. And we’ve seen a number of new concepts emerging, such as engagement, pleasure, presence, and fun, to name a few. Their proponents suggest that these concepts represent important aspects of usage that are omitted by traditional approaches to evaluation. We’ve heard it all before, of course, but the words are new. However, does the emperor actually have new clothes? In their efforts to explore concepts related to the user experience, researchers have been slow to articulate how their proposals should be measured or indeed if they can be measured at all. Clearly what is needed is an organized discussion on what aspects of use need to be investigated when the time comes for evaluation and how they could be assessed. We propose that when evaluating technology, there are three primary elements that need to be considered, namely, the product, the interaction between the user and the product, and the experience of using the product. Each of these three elements represents a unique but interdependent aspect of usage. These are Functionality (product), Usability (interaction), and Experience (user experience). Each area asks a different question about usage using a different language of discourse.

Abstract: We show how cultural selection for learnability during the process of linguistic evolution can be visualized using a simple iterated learning model. Computational models of linguistic evolution typically focus on the nature of, and conditions for, stable states. We take a novel approach and focus on understanding the process of linguistic evolution itself. What kind of evolutionary system is this process? Using visualization techniques, we explore the nature of replicators in linguistic evolution, and argue that replicators correspond to local regions of regularity in the mapping between meaning and signals. Based on this argument, we draw parallels between phenomena observed in the model and linguistic phenomena observed across languages. We then go on to identify issues of replication and selection as key points of divergence in the parallels between the processes of linguistic evolution and biological evolution.

Abstract: A sequence of prices and demands are rationalizable if there exists a concave, continuous and monotone utility function such that the demands are the maximizers of the utility function over the budget set corresponding to the price. Afriat [1] presented necessary and sufficient conditions for a finite sequence to be rationalizable. Varian [20] and later Blundell et al. [3, 4] continued this line of work studying nonparametric methods to forecasts demand. Their results essentially characterize learnability of degenerate classes of demand functions and therefore fall short of giving a general degree of confidence in the forecast. The present paper complements this line of research by introducing a statistical model and a measure of complexity through which we are able to study the learnability of classes of demand functions and derive a degree of confidence in the forecasts.Our results show that the class of all demand functions has unbounded complexity and therefore is not learnable, but that there exist interesting and potentially useful classes that are learnable from finite samples. We also present a learning algorithm that is an adaptation of a new proof of Afriat's theorem due to Teo and Vohra [17].

Abstract: We study how determinacy and learnability of global rational expectations equilibrium may be affected by monetary policy in a simple, two country, new Keynesian framework. The two blocks may be viewed as the U.S. and Europe, or as regions within the euro zone. We seek to understand how monetary policy choices may interact across borders to help or hinder the creation of a unique rational expectations equilibrium worldwide which can be learned by market participants. We study cases in which optimal policies are being pursued country by country as well as some forms of cooperation. We find that open economy considerations may alter conditions for determinacy and learnability relative to closed economy analyses, and that new concerns can arise in the analysis of classic topics such as the desirability of exchange rate targeting and monetary policy cooperation.

Abstract: The study investigates the relationship between input, UG (Universal Grammar) parameter values, and the native language in the acquisition of a purely semantic property that is superficially unrelated to its syntactic trigger, The Bare Noun/Proper Name parameter (Longobardi, 1991; 1994; 1996; 2001; 2005). On the one hand, English and Italian bare nouns have identical syntactic form and distribution, but differ in available interpretations. On the other hand, proper names display cross-linguistic constant meaning but variable word order. Variation in this respect can be accounted for by a parameter that is set to one value in English and another one in Italian. A bidirectional study of the two properties was conducted. Individual results calculated with the native speaker accuracy as the cut-off point for successful acquisition indicate that parametric restructuring is attested in both learning directions. In the English [.arrowright] Italian direction, the lack of one native interpretation in the target language (a contracting of the grammar) is achieved in the absence of negative evidence, in a Poverty of the Stimulus situation. In both directions, the semantic property is acquired based on input and/or positive evidence for the syntactic side of the parameter.

Abstract: We present a new 4-key text entry method that, unlike most few-key methods, is gestural instead of selection-based. Importantly, its gestures mimic the writing of Roman letters for high learnability. We compare this new 4-key method to predominant 3-key and 5-key methods theoretically using KSPC and empirically using a longitudinal study of 5 subjects over 10 sessions. The study includes an evaluation of the 4-key method without any on-screen visualization-an impossible condition for the selection-based methods. Our results show that the new 4-key method is quickly learned, becoming faster than the 3-key and 5-key methods after just ~10 minutes of writing, although it produces more errors. Interestingly, removing a visualization of the gestures being made causes no detriment to the 4-key method, which is an advantage for eyes-free text entry.

Abstract: Using an experimental method, we found that customers, partnered with an IT professional, are able to use executable acceptance test (storytest)-based specifications to communicate and validate functional business requirements. However, learnability and ease of use analysis indicates that an average customer may experience difficulties learning the technique. Several additional propositions are evaluated and usage observations made.

Abstract: We study how determinacy and learnability of worldwide rational expectations equilibrium may be affected by monetary policy in a simple, two-country, New Keynesian framework under both fixed and flexible exchange rates. We find that open economy considerations may alter conditions for determinacy and learnability relative to closed economy analyses and that new concerns can arise in the analysis of classic topics such as the desirability of exchange rate targeting and monetary policy cooperation.

Abstract: Virtually all generative accounts of language development assume that syntactic acquisition is guided by a learnability constraint called the Subset Principle (SP). In essence, SP forces learners to initially select the value of a parameter that generates the smallest possible language. Recent work on the acquisition of semantics suggests that there also are semantic subset problems; therefore, one needs a Semantic Subset Principle (SSP) to solve these problems. This article shows that (a) to the extent that semantic subset problems exist, the SSP is not the correct solution to them and (b) that semantic subset problems most likely do not exist in the first place. These conclusions have important implications for theories of the acquisition of semantic knowledge and for the study of language acquisition more generally. First, they provide a basis for delimiting the class of problems that theories of the acquisition of semantics ought to be responsible for. Second, these conclusions underscore the fact tha...

Abstract: Title of Document: MULTIPLE INTERROGATIVES: SYNTAX, SEMANTICS, AND LEARNABILITY Lydia Grebenyova, Ph.D., 2006 Directed By: Distinguished University Professor, Howard Lasnik, Department of Linguistics The dissertation consists of theoretical and experimental studies of multiple interrogatives (i.e., sentences containing more than one wh-phrase, like Who bought what?). First, I examine the status of Superiority effects in contexts with and without subject-aux(iliary) inversion cross-linguistically. The relevant contrast from English is between Who bought what?, What did who buy?, and *I wonder what who bought., where (*) indicates a greater degree of unacceptability by native speakers than (). I argue that the presence of subject-aux inversion in main clauses in English is responsible for the given asymmetry, and I attribute the degraded status of What did who buy? to the independent semantic properties of questions. Next, I explore the semantic properties of multiple interrogatives in detail. I develop an analysis that does not rely on covert wh-movement, relying instead on the syntactic position of the Question morpheme. I also explore the nature of complex wh-phrases (e.g., what boy, which book). I propose that choice functions are part of complex wh-phrases but not bare wh-phrases. I then explore the behavior of multiple interrogatives under Sluicing (i.e., clausal ellipsis). I observe that, in Slavic, it is possible to have multiple wh-phrases as well as focused referential expressions as remnants of sluicing. Based on this data, I argue that clausal ellipsis is licensed by focus in general. I also explore the apparent Superiority effects under sluicing in Russian and Polish and conclude that those are, in fact, parallelism effects, and not minimality effects. Finally, I present the results of several language acquisition studies on at what age and how English-, Russian-, and Malayalamspeaking children acquire the language-specific syntactic and semantic properties of multiple interrogatives, given the limited evidence in the input. I report the results of the corpus studies of parental speech with respect to the frequency of occurrence of multiple interrogatives, as well as the results of the studies, where multiple interrogatives were elicited from children and adults in specific contexts. I conclude that young children acquire syntax and semantics of multiple interrogatives quite successfully. I then discuss what evidence in the input they might be using. MULTIPLE INTERROGATIVES: SYNTAX, SEMANTICS, AND LEARNABILITY

Abstract: This paper provides a tutorial on the selection of complex target sounds for treatment following from known principles of language learnability. The focus is on syllable structure in recommending onset consonant clusters for treatment. The step-by-step procedure of cluster selection is illustrated for one child, Jarrod, who presented with a phonological disorder. Target selection procedures are guided by universal principles that govern the phonotactics of onset clusters and experimental evidence that supports the efficacy of phonologically complex targets. The prediction is that treatment of onset clusters will facilitate Jarrod's learning of both complex and simple properties of the sound system.

Abstract: Strings can be mapped into Hilbert spaces using feature maps such as the Parikh map. Languages can then be defined as the pre-image of hyperplanes in the feature space, rather than using grammars or automata. These are the planar languages. In this paper we show that using techniques from kernel-based learning, we can represent and efficiently learn, from positive data alone, various linguistically interesting context-sensitive languages. In particular we show that the cross-serial dependencies in Swiss German, that established the non-context-freeness of natural language, are learnable using a standard kernel. We demonstrate the polynomial-time identifiability in the limit of these classes, and discuss some language theoretic properties of these classes, and their relationship to the choice of kernel/feature map.

Abstract: We report six experiments on learnability of four non-adjacent phonotactic constraints which differ in their attested frequency and phonetic conditioning factors; liquid harmony, liquid disharmony, backness harmony, and backness disharmony. Our results suggest that such phonotactic constraints can be implicitly learned from brief experience and that learnability of a phonological grammar may be independent of its attested frequency and phonetic basis.

Abstract: The evolution of the World Wide Web has encouraged a huge surfacing of e-learning technologies over recent years. Often, such technology is rolled out devoid of consideration towards the way in which students process and assimilate information. To date, there exists inconclusive and contradictory evidence concerning learnability effects of single- and dual-model systems in education. To overcome this, we advocate that the design of e-learning systems requires a managed mix of elements grounded in cognitive psychology. In this paper, we report the results of a study concerned with determining an effective mix of multimedia elements. This is with regard to the situated learnability effects of single- and dual-modal systems tested via 'text only' and 'text and auditory-verbal' conditions. We report on these experiments using science computer-assisted teaching and music-oriented learning environment, e-learning environments developed to act as test platforms for this research. The results indicated that simultaneous presentation of identical information via text and narration was associated with enhanced learnability.

Abstract: We study how determinacy and learnability of worldwide rational expectations equilibrium may be affected by monetary policy in a simple, two country, New Keynesian framework under both fixed and flexible exchange rates. We find that open economy considerations may alter conditions for determinacy and learnability relative to closed economy analyses, and that new concerns can arise in the analysis of classic topics such as the desirability of exchange rate targeting and monetary policy cooperation.

Abstract: A classical approach in multi-class pattern classification is the following. Estimate the probability distributions that generated the observations for each label class, and then label new instances by applying the Bayes classifier to the estimated distributions. That approach provides more useful information than just a class label; it also provides estimates of the conditional distribution of class labels, in situations where there is class overlap. We would like to know whether it is harder to build accurate classifiers via this approach, than by techniques that may process all data with distinct labels together. In this paper we make that question precise by considering it in the context of PAC learnability. We propose two restrictions on the PAC learning framework that are intended to correspond with the above approach, and consider their relationship with standard PAC learning. Our main restriction of interest leads to some interesting algorithms that show that the restriction is not stronger (more restrictive) than various other well-known restrictions on PAC learning. An alternative slightly milder restriction turns out to be almost equivalent to unrestricted PAC learning.

Abstract: We address the issue of the learnability of concept classes under three classification noise models in the probably approximately correct framework. After introducing the Class-Conditional Classification Noise (CCCN) model, we investigate the problem of the learnability of concept classes under this particular setting and we show that concept classes that are learnable under the well-known uniform classification noise (CN) setting are also CCCN-learnable, which gives CN = CCCN. We then use this result to prove the equality between the set of concept classes that are CN-learnable and the set of concept classes that are learnable in the Constant Partition Classification Noise (CPCN) setting, or, in other words, we show that CN = CPCN.

Abstract: Inductive inference can be considered as one of the fundamental paradigms of algorithmic learning theory. We survey results recently obtained and show their impact to potential applications.Since the main focus is put on the efficiency of learning, we also deal with postulates of naturalness and their impact to the efficiency of limit learners. In particular, we look at the learnability of the class of all pattern languages and ask whether or not one can design a learner within the paradigm of learning in the limit that is nevertheless efficient.For achieving this goal, we deal with iterative learning and its interplay with the hypothesis spaces allowed. This interplay has also a severe impact to postulates of naturalness satisfiable by any learner.Furthermore, since a limit learner is only supposed to converge, one never knows at any particular learning stage whether or not the learner did already succeed. The resulting uncertainty may be prohibitive in many applications. We survey results to resolve this problem by outlining a new learning model, called stochastic finite learning. Though pattern languages can neither be finitely inferred from positive data nor PAC-learned, our approach can be extended to a stochastic finite learner that exactly infers all pattern languages from positive data with high confidence.Finally, we apply the techniques developed to the problem of learning conjunctive concepts.

Abstract: A model for learning in the limit is defined where a (so-called iterative) learner gets all positive examples from the target language, tests every new conjecture with a teacher (oracle) if it is a subset of the target language (and if it is not, then it receives a negative counterexample), and uses only limited long-term memory (incorporated in conjectures). Three variants of this model are compared: when a learner receives least negative counterexamples, the ones whose size is bounded by the maximum size of input seen so far, and arbitrary ones. We also compare our learnability model with other relevant models of learnability in the limit, study how our model works for indexed classes of recursive languages, and show that learners in our model can work in non-U-shaped way — never abandoning the first right conjecture.

Abstract: Background Several studies indicate that EEG alpha activity is associated with cognitive performance and learnability (Anokhin and Vogel, 1996, Klimesch and Doppelmayr, 1998). It was demonstrated that resting alpha power is increased under conditions that are associated with enhanced cognitive processing capacity or situations where subjects try to increase their capacity. (Klimesch, 1999) Different parameters of alpha, however, are related to different aspects of cognitive performance and learnability in different ways

Abstract: We address the problem of reinforcement learning in which observations may exhibit an arbitrary form of stochastic dependence on past observations and actions, i.e. environments more general than (PO) MDPs. The task for an agent is to attain the best possible asymptotic reward where the true generating environment is unknown but belongs to a known countable family of environments. We find some sufficient conditions on the class of environments under which an agent exists which attains the best asymptotic reward for any environment in the class. We analyze how tight these conditions are and how they relate to different probabilistic assumptions known in reinforcement learning and related fields, such as Markov Decision Processes and mixing conditions.

Abstract: We study the problem of learning an unknown function represented as an expression or a program over a known finite monoid As in other areas of computational complexity where programs over algebras have been used, the goal is to relate the computational complexity of the learning problem with the algebraic complexity of the finite monoid Indeed, our results indicate a close connection between both kinds of complexity We focus on monoids which are either groups or aperiodic, and on the learning model of exact learning from queries For a group G, we prove that expressions over G are efficiently learnable if G is nilpotent, and impossible to learn efficiently (under cryptographic assumptions) if G is nonsolvable We present some results for restricted classes of solvable groups, and point out a connection between their efficient learnability and the existence of lower bounds on their computational power in the program model For aperiodic monoids, our results seem to indicate that the monoid class known as DA captures exactly learnability of expressions by polynomially many Evaluation queries When using programs instead of expressions, we show that our results for groups remain true, while the situation is quite different for aperiodic monoids

Abstract: ion, encapsulation, lazy evaluation, loaded symbols polymorphism in Smalltalk -you send the message print, and each receiver has its own version (implementation) of print: A message specifies which operation is designed, but not how that operation should be carried out (Goldberg & Robson, 1983) Goldberg, Adele E. & David Robson. Smalltalk-80: the language and its implementation. Addison Wesley, 1983. 39 Information and Communication Technologies 5. Language evolution origins (of underlying mechanisms) emergence (of specific features) evolution proper Vogt, Paul, Bart de Boer & Tony Belpaeme. “Modelling language origins and evolution”, Tutorial, IJCAI 2005 (31 July 2005, Edinburgh, Scotland). evolution of communication grounding (relating language to the world) computational simulation, the emergence of compositionality 40 Information and Communication Technologies Examples of language evolution viuvar -> enviuvar conheco -> reconheco costumar-se -> acostumar-se amar hei-de -> amarei cases in Latin -> articles in Romance languages amara -> tinha amado buscar -> ir buscar jogar a -> jogar a amar a -> amar 41 Information and Communication Technologies Programming language evolution versioning deprecated syntax documentation compilers’ warning messages upwards compatibility solving errors in previous versions different language paradigms (imperative, functional, ...) object orientation, extreme programming, ... Difference between language designers and speakers 42 Information and Communication Technologies Acquisition and learnability of language In order to learn language, one has to use it: thus the primacy of dialogue and of context Every generation learns it anew... In order to learn, one has to actively extend it and see whether the extensions are sanctioned or not by future dialogues (Sampson’s little Popperian in Educating Eve) Vagueness is a tool to help learning: not more than necessary needs to be decided/conveyed in an interaction, things can be refined, if needed, in dialogue

Abstract: This thesis considers probabilistic finite automata (PFA), distributions of sequences over finite alphabets, the links between them and the learnability thereof. Pervasive in scientific fields ranging from computer science to electrical engineering to information theory, PFA models also find numerous practical applications in speech recognition, bioinformatics and natural language processing. PFA models are the most general among the myriad of syntactic objects providing probabilistic extensions of finite state machines. Closely related to hidden Markov models (HMMs), PFAs have been the focus of extensive research, but continue to pose interesting theoretical and practical problems to this day. The thesis presents geometric insights into the PFA learning problem, a characterization theorem for the family of distributions induced by PFA models, as well as a number of applications of this theorem. For a subclass of PFA called probabilistic deterministic finite automata (PDFA), a number of learnability results are presented. These results place limits on the PDFA subclasses which are learnable using a class of algorithms collectively known as state merging. The sample complexity of learning general distributions over countable sets is considered, and lower and upper bounds, which asymptotically match up to a logarithmic factor are developed. An example is constructed exhibiting a class of PDFA models which is efficiently learnable using state merging. It is demonstrated that distributions induced by this class are not efficiently learnable by direct estimation (making no assumptions on the distribution’s source) in the sense that the sample complexity is bounded below by an exponential in the number of states.

Abstract: Research on learning has shown that student learn more effectively when taught with methods that suits to their learning style. Regarding to the problem, an adaptive hypermedia learning system that exploit a mixed approach has been developed. The mixed approach comprise of 2 approaches; computer intelligence (fuzzy logic approach) and personality factor (MBTI approach) that use to individualize the learning material structure. It aims to utilize the learning characteristics and provide a personalized learning environment that exploit pedagogical model and fuzzy logic techniques. The learning material consists of 4 structures; 1) theory, 2) examples 3) exercises and 4) activities. The pedagogical model and learning characteristics are based on the student’s personality factor (Myers-Briggs Type Indicator (MBTI)), whilst the fuzzy logic techniques are used to classify the structure of learning material which is based on student’s personality factors. This paper tends to exemplify the evaluation process for mixed approach. The evaluation is comprised of two methods; usability and utility which is both of the methods are referring to the learnability factor, efficiency factor, satisfaction factor and accuracy factor. The system has been tested by 44 students from Faculty of Computer Science. Most of the participants give good response and the results from the evaluation have been very satisfactory.

Abstract: Recently, Kalai et al. [1] have shown (among other things) that linear threshold functions over the Boolean cube and unit sphere are agnostically learnable with respect to the uniform distribution using the hypothesis class of polynomial threshold functions. Their primary algorithm computes monomials of large constant degree, although they also analyze a low-degree algorithm for learning origin-centered halfspaces over the unit sphere. This paper explores noise-tolerant learnability of linear thresholds over the cube when the learner sees a very limited portion of each instance. Uniform-distribution weak learnability results are derived for the agnostic, unknown attribute noise, and malicious noise models. The noise rates that can be tolerated vary: the rate is essentially optimal for attribute noise, constant (roughly 1/8) for agnostic learning, and non-trivial ($\Omega(1/\sqrt{n})$) for malicious noise. In addition, a new model that lies between the product attribute and malicious noise models is introduced, and in this stronger model results similar to those for the standard attribute noise model are obtained for learning homogeneous linear thresholds with respect to the uniform distribution over the cube. The learning algorithms presented are simple and have small-polynomial running times.

Abstract: A brief timetable of language development is outlined. This article then deals in turn with the major aspects of language learning in terms of infancy, learning words, learning morphology, early grammar, later grammar, the learning of pragmatic and metalinguistic skills and, finally, some brief reflections on atypical development. The relevant theoretical issues are covered as they arise in each section. They are taken up again in the last section on learnability and constituency.