Page's proposal to stipulate representations in which individual units correspond to meaningful entities is too unconstrained to support effective theorizing. An approach combining general computational principles with domain-specific assumptions, in which learning is used to discover representations that are effective in solving tasks, provides more insight into why cognitive and neural systems are organized the way they are.

We agree with the authors that putting forward specific models and examining their agreement with experimental data are the best approach for understanding the nature of decision making. Although the authors only consider the likelihood function, prior, cost function, and decision rule (LPCD) framework, other choices are available. Bayesian statistics can be used to estimate essential parameters and assess the degree of optimality.

The discovery of the first electromagnetic counterpart to a gravitational wave signal has generated follow-up observations by over 50 facilities world-wide, ushering in the new era of multi-messenger astronomy. In this paper, we present follow-up observations of the gravitational wave event GW170817 and its electromagnetic counterpart SSS17a/DLT17ck (IAU label AT2017gfo) by 14 Australian telescopes and partner observatories as part of Australian-based and Australian-led research programs. We report early- to late-time multi-wavelength observations, including optical imaging and spectroscopy, mid-infrared imaging, radio imaging, and searches for fast radio bursts. Our optical spectra reveal that the transient source emission cooled from approximately 6 400 K to 2 100 K over a 7-d period and produced no significant optical emission lines. The spectral profiles, cooling rate, and photometric light curves are consistent with the expected outburst and subsequent processes of a binary neutron star merger. Star formation in the host galaxy probably ceased at least a Gyr ago, although there is evidence for a galaxy merger. Binary pulsars with short (100 Myr) decay times are therefore unlikely progenitors, but pulsars like PSR B1534+12 with its 2.7 Gyr coalescence time could produce such a merger. The displacement (~2.2 kpc) of the binary star system from the centre of the main galaxy is not unusual for stars in the host galaxy or stars originating in the merging galaxy, and therefore any constraints on the kick velocity imparted to the progenitor are poor.

Lake et al. propose that people rely on “start-up software,” “causal models,” and “intuitive theories” built using compositional representations to learn new tasks more efficiently than some deep neural network models. We highlight the many drawbacks of a commitment to compositional representations and describe our continuing effort to explore how the ability to build on prior knowledge and to learn new tasks efficiently could arise through learning in deep neural networks.

Studies were conducted to determine injury potential to rotational crops from carryover of herbicides used in watermelon production. Treatments included halosulfuron, ethalfluralin, and sulfentrazone alone; halosulfuron in tank mixtures with bensulide, clomazone, ethalfluralin, and naptalam; and a tank mixture of naptalam and bensulide. Sulfentrazone applied at 224 g ai/ha to watermelon severely reduced spinach emergence, but did not reduce emergence of broccoli, cabbage, or wheat. Residues of sulfentrazone applied to watermelon at 450 g/ha stunted growth of broccoli and cabbage and was the only treatment that reduced wheat stand. Injury to broccoli, cabbage, and spinach increased as the halosulfuron rate increased. Ethalfluralin did not reduce stand or cause injury to any of the four rotational crops. Naptalam plus bensulide did not reduce stand of the four crops and caused either slight or no injury. Residues of sulfentrazone and halosulfuron can injure vegetables following crops in which these herbicides are used, and caution should be taken particularly with spinach, broccoli, and cabbage in this respect.

Studies were conducted at eight sites during a 3-yr period in Oklahoma and Arkansas to determine the effectiveness and safety of preemergence applications of halosulfuron both alone and in tank mixtures with bensulide, clomazone, ethalfluralin, and naptalam. Ethalfluralin, naptalam plus bensulide, and sulfentrazone also were applied alone. Although halosulfuron caused up to 20% seedling stunting, watermelon plants recovered by 5 to 7 wk after planting, and yield was similar to that of hand-weeded plots. Halosulfuron treatments controlled hophornbeam copperleaf, Palmer amaranth, carpetweed, and cutleaf groundcherry 80 to 100%. Control of goosegrass was at least 97% with clomazone plus ethalfluralin plus halosulfuron. Injury to watermelon treated with sulfentrazone ranged from 76 to 98% at 2 to 4 wk after treatment. This was reflected by yields that were lower than any other herbicide treatment in the studies.

Many attempts have been made to teach native Japanese listeners to perceptually differentiate English /r–l/ (e.g. rock–lock). Though improvement is evident, in no case is final performance native English-like. We focused our training on the third formant onset frequency, shown to be the most reliable indicator of /r–l/ category membership. We first presented listeners with instances of synthetic /r–l/ stimuli varying only in F3 onset frequency, in a forced-choice identification training task with feedback. Evidence of learning was limited. The second experiment utilized an adaptive paradigm beginning with non-speech stimuli consisting only of /r/ and /l/ F3 frequency trajectories progressing to synthetic speech instances of /ra–la/; half of the trainees received feedback. Improvement was shown by some listeners, suggesting some enhancement of /r–l/ identification is possible following training with only F3 onset frequency. However, only a subset of these listeners showed signs of generalization of the training effect beyond the trained synthetic context.

Mitchell et al. describe many fascinating studies, and in the process, propose what they consider to be a unified framework for human learning in which effortful, controlled learning results in propositional knowledge. However, it is unclear how any of their findings privilege a propositional account, and we remain concerned that embedding all knowledge in propositional representations obscures the tight interdependence between learning from experiences and the use of the results of learning as a basis for action.

In this précis of our recent book, Semantic Cognition: A Parallel Distributed Processing Approach (Rogers & McClelland 2004), we present a parallel distributed processing theory of the acquisition, representation, and use of human semantic knowledge. The theory proposes that semantic abilities arise from the flow of activation among simple, neuron-like processing units, as governed by the strengths of interconnecting weights; and that acquisition of new semantic information involves the gradual adjustment of weights in the system in response to experience. These simple ideas explain a wide range of empirical phenomena from studies of categorization, lexical acquisition, and disordered semantic cognition. In this précis we focus on phenomena central to the reaction against similarity-based theories that arose in the 1980s and that subsequently motivated the “theory-theory” approach to semantic knowledge. Specifically, we consider (1) how concepts differentiate in early development, (2) why some groupings of items seem to form “good” or coherent categories while others do not, (3) why different properties seem central or important to different concepts, (4) why children and adults sometimes attest to beliefs that seem to contradict their direct experience, (5) how concepts reorganize between the ages of 4 and 10, and (6) the relationship between causal knowledge and semantic knowledge. The explanations our theory offers for these phenomena are illustrated with reference to a simple feed-forward connectionist model. The relationships between this simple model, the broader theory, and more general issues in cognitive science are discussed.

The commentaries reflect three core themes that pertain not just to our theory, but to the enterprise of connectionist modeling more generally. The first concerns the relationship between a cognitive theory and an implemented computer model. Specifically, how does one determine, when a model departs from the theory it exemplifies, whether the departure is a useful simplification or a critical flaw? We argue that the answer to this question depends partially upon the model's intended function, and we suggest that connectionist models have important functions beyond the commonly accepted goals of fitting data and making predictions. The second theme concerns perceived in-principle limitations of the connectionist approach to cognition, and the specific concerns these perceived limitations raise for our theory. We argue that the approach is not in fact limited in the ways our critics suggest. One common misconception, that connectionist models cannot address abstract or relational structure, is corrected through new simulations showing directly that such structure can be captured. The third theme concerns the relationship between parallel distributed processing (PDP) models and structured probabilistic approaches. In this case we argue that there the difference between the approaches is not merely one of levels. Our PDP approach differs from structured statistical approaches at all of Marr's levels, including the characterization of the goals of cognitive computations, and of the representations and algorithms used.

We share with Anderson & Lebiere (A&L) (and with Newell before them) the goal of developing a domain-general framework for modeling cognition, and we take seriously the issue of evaluation criteria. We advocate a more focused approach than the one reflected in Newell's criteria, based on analysis of failures as well as successes of models brought into close contact with experimental data. A&L attribute the shortcomings of our parallel-distributed processing framework to a failure to acknowledge a symbolic level of thought. Our framework does acknowledge a symbolic level, contrary to their claim. What we deny is that the symbolic level is the level at which the principles of cognitive processing should be formulated. Models cast at a symbolic level are sometimes useful as high-level approximations of the underlying mechanisms of thought. The adequacy of this approximation will continue to increase as symbolic modelers continue to incorporate principles of parallel distributed processing.

Different developmental anomalies produce contrasting deficits in a single, integrated system. In a network that inflects regular and exception verbs correctly, a disproportionate deficit with exceptions occurs if connections are deleted, whereas a disproportionate deficit with regulars occurs when an auditory deficit impairs perception of the regular inflection. In general, contrasting deficits do not license the inference of underlying modularity.