Brainstorming suggests that generating ideas in groups can direct one's thoughts into areas that might be difficult to reach or consider, resulting in better performance than individuals working separately. However, while intuitively appealing, studies have repeatedly found no evidence for such synergy. Here, we reinvestigate group synergy in a task with simulated agents. We use state-of-the-art language embeddings to generate close or far-off semantic associates to people's ideas. This design allows us to adopt a sampling for inference perspective, where individuals produce sample ideas from a generative process (Sanborn & Chater, 2016). We hypothesize that suggestions can allow participants to create samples from previously not considered concepts, and additionally, knowledge about the diversity of solutions can spur more broad idea generation. Our work helps inform broader questions about when and why groups can expect to produce better performance than individuals.

Human probability judgements appear systematically biased, in apparent tension with Bayesian models of cognition. Butperhaps the brain does not represent probabilities explicitly, but approximates probabilistic calculations through a processof sampling, as used in computational probabilistic models in statistics. The Bayesian sampling viewpoint provides asimple rational model of probability judgements, which generates biases such as conservatism. The Bayesian samplerprovides a single framework for explaining phenomena associated with diverse biases and heuristics, including availabilityand representativeness. The approach turns out to provide a rational reinterpretation of noise in an important recent modelof probability judgement, the probability theory plus noise model (Costello & Watts, 2014; 2016; 2017; Costello, Watts,& Fisher, 2018), and captures the empirical data supporting this model.

Many models of cognition assume that people can generate independent samples, yet people fail to do so in random generation tasks. One prominent explanation for this behavior is that people use learned schemas. Instead, we propose that deviations from randomness arise from people sampling locally rather than independently. To test these explanations, we teach people one- and two-dimensional arrangements of syllables and ask them to generate random sequences from them. Although our results reproduce characteristic features of human random generation, such as a preference for adjacent items and an avoidance of repetitions, we also find an effect of dimensionality on the patterns people produce. Furthermore, model comparisons revealed that local sampling accounted better for participants' sequences than a schema account. Finally, evaluating the importance of each models' constituents, we show that the local sampling model proposed new states based on its current trajectory, rather than an inhibition-of-return-like principle.

Sampling-based models, which assume that people remember or simulate instances of events and count outcome propor-tions to make probability judgments, account for many apparent biases in human judgment. The success of such modelsis generally dependent on sample size, as particularly large or particularly small samples are often required for a modelto reproduce effects observed in data. Thus, systematically exploring the actual number of instances that people tend tosample is an important criterion in evaluating model credibility. Here we propose a method of estimating sample size byway of inter-judgment variance. We show through model recovery that this method will reliably recover the correct samplesize and subsequently apply the method to two well-supported models of human probability judgment, Probability TheoryPlus Noise and the Bayesian Sampler. Results indicate, in both cases, that human probability judgments are based on arelatively plausible (¡ 10) number of sampled instances.