Fourier transform infrared (FTIR) spectrometry is commonly used for the identification of reference substances (RSs) in solid, liquid, or gaseous mixtures. An expert is generally required to perform the analysis, which is a bottleneck in emergency situations. This study proposes a support vector machine (SVM)‐based algorithm, the peak correlation classifier (PCC), designed to rapidly detect the presence of a specific threat or reference substance in a sample. While SVM has been used in various spectrographic contexts, it has rarely been used on FTIR spectra. The proposed algorithm discovers correlation similarities between the FTIR spectrum of the RS and the test sample and then uses SVM to determine whether or not the RS is present in the sample. The study also shows how the additive nature of FTIR spectra can be used to create ‘synthetic’ substances that significantly improve the detection capability and decision confidence of the SVM classifier.

Being able to automatically perceive a variety of emotions from text alone has potentially important applications in CMC and HCI that range from identifying mood from online posts to enabling dynamically adaptive interfaces. However, such ability has not been proven in human raters or computational systems. Here we examine the ability of naive raters of emotion to detect one of eight emotional categories from 50 and 200 word samples of real blog text. Using expert raters as a 'gold standard', naive-expert rater agreement increased with longer texts, and was high for ratings of joy, disgust, anger and anticipation, but low for acceptance and 'neutral' texts. We discuss these findings in light of theories of CMC and potential applications in HCI.

Emotion is central to human interactions, and automatic detection could enhance our experience with technologies. We investigate the linguistic expression of fine-grained emotion in 50 and 200 word samples of real blog texts previously coded by expert and naive raters. Content analysis (LIWC) reveals angry authors use more affective language and negative affect words, and that joyful authors use more positive affect words. Additionally, a co-occurrence semantic space approach (LSA) was able to identify fear (which naive human emotion raters could not do). We relate our findings to human emotion perception and note potential computational applications.

While humans forget gradually, highly distributed connectionist networks forget catastrophically: newly learned information often completely erases previously learned information. This is not just implausible cognitively, but disastrous practically. However, it is not easy in connectionist cognitive modelling to keep away from highly distributed neural networks, if only because of their ability to generalize. A realistic and effective system that solves the problem of catastrophic interference in sequential learning of ‘static’ (i.e. non-temporally ordered) patterns has been proposed recently (Robins 1995, Connection Science, 7: 123–146, 1996, Connection Science, 8: 259–275, Ans and Rousset 1997, CR Académie des Sciences Paris, Life Sciences, 320: 989–997, French 1997, Connection Science, 9: 353–379, 1999, Trends in Cognitive Sciences, 3: 128–135, Ans and Rousset 2000, Connection Science, 12: 1–19). The basic principle is to learn new external patterns interleaved with internally generated ‘pseudopatterns’ (generated from random activation) that reflect the previously learned information. However, to be credible, this self-refreshing mechanism for static learning has to encompass our human ability to learn serially many temporal sequences of patterns without catastrophic forgetting. Temporal sequence learning is arguably more important than static pattern learning in the real world. In this paper, we develop a dual-network architecture in which self-generated pseudopatterns reflect (non-temporally) all the sequences of temporally ordered items previously learned. Using these pseudopatterns, several self-refreshing mechanisms that eliminate catastrophic forgetting in sequence learning are described and their efficiency is demonstrated through simulations. Finally, an experiment is presented that evidences a close similarity between human and simulated behaviour.

In error-driven distributed feedforward networks, new information typically interferes, sometimes severely, with previously learned information. We show how noise can be used to approximate the error surface of previously learned information. By combining this approximated error surface with the error surface associated with the new information to be learned, the network's retention of previously learned items can be improved and catastrophic interference significantly reduced. Further, we show that the noise-generated error surface is produced using only first-derivative information and without recourse to any explicit error information.

Turney (2001) claims that a simple program, PMI-IR, that searches the World Wide Web for co-occurrences of words in 350 million Web pages can be used to find human-like answers to the type of ‘subcognitive’ questions French (1990) claimed would invariably unmask computers (that had not lived life as we humans had) in a Turing Test. This paper shows that there are serious problems with Turney's claim. We show that PMI-IR does not work for even simple subcognitive questions. PMI-IR's failure is attributed to its inability to understand the relational and contextual attributes of the words/concepts in the queries. Finally, it is shown that, even if PMI-IR were able to answer many subcognitive questions, a clever interrogator in the Turing Test would still be able to unmask the computer.

No computer that had not experienced the world as we humans had could pass a rigorously administered standard Turing Test. This paper will show that the use of ‘subcognitive’ questions allows the standard Turing Test to indirectly probe the human subcognitive associative concept network built up over a lifetime of experience with the world. Not only can this probing reveal differences in cognitive abilities, but crucially, even differences in physical aspects of the candidates can be detected. Consequently, it is unnecessary to propose even harder versions of the Test in which all physical and behavioural aspects of the two candidates had to be indistinguishable before allowing the machine to pass the Test. Any machine that passed the ‘simpler’ symbols-in symbols-out test as originally proposed by Turing would be intelligent. The problem is that, even in its original form, the Turing Test is already too hard and too anthropocentric for any machine that was not a physical, social and behavioural carbon copy of ourselves to actually pass it. Consequently, the Turing Test, even in its standard version, is not a reasonable test for general machine intelligence. There is no need for an even stronger version of the Test.

In order to solve the 'sensitivity-stability' problem-and its immediate correlate, the problem of sequential learning-it is crucial to develop connectionist architectures that are simultaneously sensitive to, but not excessively disrupted by, new input. French (1992) suggested that to alleviate a particularly severe form of this disruption, catastrophic forgetting, it was necessary for networks to separate dynamically their internal representations during learning. McClelland et al. (1995) went even further. They suggested that nature's way of implementing this obligatory separation was the evolution of two separate areas of the brain, the hippocampus and the neocortex. In keeping with this idea of radical separation, a 'pseudo-recurrent' memory model is presented here that partitions a connectionist network into two functionally distinct, but continually interacting areas. One area serves as a final-storage area for representations; the other is an early-processing area where new representations are first learned by the system. The final-storage area continually supplies internally generated patterns (pseudopatterns; Robins, 1995), which are approximations of its content, to the early-processing area, where they are interleaved with the new patterns to be learned. Transfer of the new learning is done either by weight-copying from the early-processing area to the final-storage area or by pseudopattern transfer. A number of experiments are presented that demonstrate the effectiveness of this approach, allowing, in particular, effective sequential learning with gradual forgetting in the presence of new input. Finally, it is shown that the two interacting areas automatically produce representational compaction and it is suggested that similar representational streamlining may exist in the brain.

High-level perception—the process of making sense of complex data at an abstract, conceptual level—is fundamental to human cognition. Through high-level perception, chaotic environmental stimuli are organized into mental representations that are used throughout cognitive processing. Much work in traditional artificial intelligence has ignored the process of high-level perception, by starting with hand-coded representations. In this paper, we argue that this dismissal of perceptual processes leads to distorted models of human cognition. We examine some existing artificial-intelligence models—notably BACON, a model of scientific discovery, and the Structure-Mapping Engine, a model of analogical thought—-and argue that these are flawed precisely because they downplay the role of high-level perception. Further, we argue that perceptual processes cannot be separated from other cognitive processes even in principle,and therefore that traditional artificial-intelligence models cannot be defended by supposing the existence of a ‘representation module’ that supplies representations ready-made. Finally, we describe a model of high-level perception and analogical thought in which perceptual processing is integrated with analogical mapping, leading to the flexible build-up of representations appropriate to a given context.