Deception-detection and machine intelligence in practical Turing tests

Deception-detection is the crux of Turing’s experiment to examine machine thinking conveyed through a capacity to respond with sustained and satisfactory answers to unrestricted questions put by a human interrogator. However, in 60 years to the month since the publication of Computing Machinery and Intelligence little agreement exists for a canonical format for Turing’s textual game of imitation, deception and machine intelligence. This research raises from the trapped mine of philosophical claims, counter-claims and rebuttals Turing’s own distinct five minutes question-answer imitation game, which he envisioned practicalised in two different ways: a) A two-participant, interrogator-witness viva voce, b) A three-participant, comparison of a machine with a human both questioned simultaneously by a human interrogator. Using Loebner’s 18th Prize for Artificial Intelligence contest, and Colby et al.’s 1972 transcript analysis paradigm, this research practicalised Turing’s imitation game with over 400 human participants and 13 machines across three original experiments. Results show that, at the current state of technology, a deception rate of 8.33% was achieved by machines in 60 human-machine simultaneous comparison tests. Results also show more than 1 in 3 Reviewers succumbed to hidden interlocutor misidentification after reading transcripts from experiment 2. Deception-detection is essential to uncover the increasing number of malfeasant programmes, such as CyberLover, developed to steal identity and financially defraud users in chatrooms across the Internet. Practicalising Turing’s two tests can assist in understanding natural dialogue and mitigate the risk from cybercrime.

A reappraisal of blood clotting response tests for anticoagulant resistance and a proposal for a standardised BCR test methodology

Blood clotting response (BCR) resistance tests are available for a number of anticoagulant rodenticides. However, during the development of these tests many of the test parameters have been changed, making meaningful comparisons between results difficult. It was recognised that a standard methodology was urgently required for future BCR resistance tests and, accordingly, this document presents a reappraisal of published tests, and proposes a standard protocol for future use (see Appendix). The protocol can be used to provide information on the incidence and degree of resistance in a particular rodent population; to provide a simple comparison of resistance factors between active ingredients, thus giving clear information about cross-resistance for any given strain; and to provide comparisons of susceptibility or resistance between different populations. The methodology has a sound statistical basis in being based on the ED50 response, and requires many fewer animals than the resistance tests in current use. Most importantly, tests can be used to give a clear indication of the likely practical impact of the resistance on field efficacy. The present study was commissioned and funded by the Rodenticide Resistance Action Committee (RRAC) of CropLife International.

Thinking for speaking and linguistic relativity among bilinguals: towards a new research agenda

This article evaluates how the different papers in this special issue fill a gap in our understanding of cognitive processes that are being activated when second language learners or bilinguals prepare to speak. All papers are framed in Slobin’s (1987) Thinking for Speaking theory, and aim to test whether the conceptualisation patterns that were learned in early childhood can be relearned or restructured in L2 acquisition. In many papers the focus is on identifying constraints on this restructuring process. Among these constraints, the role of typological differences between languages is investigated in great depth. The studies involve different types of learners, language combinations and tasks. As all informants were given verbal rather than non-verbal tasks, the focus is here on the effects of conceptual transfer from one language on another, and not on the effects of language on non-linguistic cognition. The paper also sketches different avenues for further research in this field and proposes that researchers working in this field might want to take up the challenge of investigating whether speakers of different languages perceive motion outside explicitly verbal contexts differently, as this will enable us to gain an understanding of linguistic relativity effects in this domain. Studying which teaching methods can help learners to restructure their conceptualisation patterns may also shed new light on the aspects of discourse organization and motion event construal that are most difficult for learners.

Reviewing the sustainability/stationarity of current account imbalances with tests for bounded integration

We investigate for 26 OECD economies whether their current account imbalances to GDP are driven by stochastic trends. Regarding bounded stationarity as the more natural counterpart of sustainability, results from Phillips–Perron tests for unit root and bounded unit root processes are contrasted. While the former hint at stationarity of current account imbalances for 12 economies, the latter indicate bounded stationarity for only six economies. Through panel-based test statistics, current account imbalances are diagnosed as bounded non-stationary. Thus, (spurious) rejections of the unit root hypothesis might be due to the existence of bounds reflecting hidden policy controls or financial crises.

Some implications of a sample of practical Turing tests

A series of imitation games involving 3-participant (simultaneous comparison of two hidden entities) and 2-participant (direct interrogation of a hidden entity) were conducted at Bletchley Park on the 100th anniversary of Alan Turing’s birth: 23 June 2012. From the ongoing analysis of over 150 games involving (expert and non-expert, males and females, adults and child) judges, machines and hidden humans (foils for the machines), we present six particular conversations that took place between human judges and a hidden entity that produced unexpected results. From this sample we focus on features of Turing’s machine intelligence test that the mathematician/code breaker did not consider in his examination for machine thinking: the subjective nature of attributing intelligence to another mind.

The ‘thinking’ in thinking-for-speaking: where is it?

According to the thinking-for-speaking (TFS) hypothesis, speakers of different languages think differently while in the process of mentally preparing content for speech. The aim of the present paper is to critically discuss the research carried out within the TFS paradigm, against the background of the basic tenets laid out by the proponents of this framework. We will show that despite substantial progress in the investigation of crosslinguistic differences in the organisation of information in discourse, the studies that actually examine the cognitive aspects of speech production are, to date, vanishingly few. This state of affairs creates a gap in our knowledge about the thought processes that co-occur with speech production during language use and acquisition. We will argue that in order to reach a more comprehensive picture of the cognitive processes and outcomes of speech production, methodologies additional to the analysis of information organisation must be used.

Broadly speaking: vocabulary in semantic dementia shifts towards general, semantically diverse words

One of the cardinal features of semantic dementia (SD) is a steady reduction in expressive vocabulary. We investigated the nature of this breakdown by assessing the psycholinguistic characteristics of words produced spontaneously by SD patients during an autobiographical memory interview. Speech was analysed with respect to frequency and imageability, and a recently-developed measure called semantic diversity. This measure quantifies the degree to which a word can be used in a broad range of different linguistic contexts. We used this measure in a formal exploration of the tendency for SD patients to replace specific terms with more vague and general words, on the assumption that more specific words are used in a more constrained set of contexts. Relative to healthy controls, patients were less likely to produce low-frequency, high-imageability words, and more likely to produce highly frequent, abstract words. These changes in the lexical-semantic landscape were related to semantic diversity: the highly frequent and abstract words most prevalent in the patients' speech were also the most semantically diverse. In fact, when the speech samples of healthy controls were artificially engineered such that low semantic diversity words (e.g., garage, spanner) were replaced with broader terms (e.g., place, thing), the characteristics of their speech production came to closely resemble that of SD patients. A similar simulation in which low-frequency words were replaced was less successful in replicating the patient data. These findings indicate systematic biases in the deterioration of lexical-semantic space in SD. As conceptual knowledge degrades, speech increasingly consists of general terms that can be applied in a broad range of linguistic contexts and convey less specific information.

Forecast encompassing tests and probability forecasts

We consider tests of forecast encompassing for probability forecasts, for both quadratic and logarithmic scoring rules. We propose test statistics for the null of forecast encompassing, present the limiting distributions of the test statistics, and investigate the impact of estimating the forecasting models' parameters on these distributions. The small-sample performance is investigated, in terms of small numbers of forecasts and model estimation sample sizes. We show the usefulness of the tests for the evaluation of recession probability forecasts from logit models with different leading indicators as explanatory variables, and for evaluating survey-based probability forecasts.

On the transfer of prior tests or study events to subsequent study

Tests, as learning events, are often more effective than are additional study opportunities, especially when recall is tested after a long retention interval. To what degree, though, do prior test or study events support subsequent study activities? We set out to test an implication of Bjork and Bjork’s (1992) new theory of disuse—that, under some circumstances, prior study may facilitate subsequent study more than does prior testing. Participants learned English–Swahili translations and then underwent a practice phase during which some items were tested (without feedback) and other items were restudied. Although tested items were better recalled after a 1-week delay than were restudied items, this benefit did not persist after participants had the opportunity to study the items again via feedback. In fact, after this additional study opportunity, items that had been restudied earlier were better recalled than were items that had been tested earlier. These results suggest that measuring the memorial consequences of testing requires more than a single test of retention and, theoretically, a consideration of the differing status of initially recallable and nonrecallable items.

A comparison of tests of non-linear cointegration with an application to the predictability of US interest rates using the term structure

We test whether there are nonlinearities in the response of short- and long-term interest rates to the spread in interest rates, and assess the out-of-sample predictability of interest rates using linear and nonlinear models. We find strong evidence of nonlinearities in the response of interest rates to the spread. Nonlinearities are shown to result in more accurate short-horizon forecasts, especially of the spread.

Bicorrelations and cross-bicorrelations as non-linearity tests and tools for exchange rate forecasting

This paper proposes and implements a new methodology for forecasting time series, based on bicorrelations and cross-bicorrelations. It is shown that the forecasting technique arises as a natural extension of, and as a complement to, existing univariate and multivariate non-linearity tests. The formulations are essentially modified autoregressive or vector autoregressive models respectively, which can be estimated using ordinary least squares. The techniques are applied to a set of high-frequency exchange rate returns, and their out-of-sample forecasting performance is compared to that of other time series models