Multiple Object Tracking Without Pre-attentive Indexing.

Multiple object tracking (MOT) involves simultaneous tracking of a certain number of target objects amongst a larger set of objects as they all move unpredictably over time. The prevalent explanation for successful target tracking by humans in MOT involving visually identical objects is based on the Visual Indexing Theory. This assumes that each target is indexed by a pointer using a non-conceptual mechanism to maintain an object's identity even as its properties change over time. Thus, successful tracking requires successful indexing and the absence of identification errors. Identity maintenance and successful tracking are measured in terms of identification (ID) and tracking accuracy respectively, with higher accuracy indicating better identity maintenance or better tracking. Existing evidence suggests that humans have high tracking accuracy despite poor identification accuracy, suggesting that it might be possible to perform MOT without indexing. Our work adds to existing evidence for this position through two experiments, and presents a computational model of multiple object tracking that does not require indexes. Our empirical results show that identification accuracy is aligned with tracking accuracy in humans for tracking up to three, but is lower when tracking more objects. Our computational model of MOT without indexing accounts for several empirical tracking accuracy patterns shown in earlier studies, reproduces the dissociation between tracking and identification accuracy produced earlier in the literature as well as in our experiments, and makes several novel predictions.

Statistical prediction alone cannot identify good models of behavior.

The dissociation between statistical prediction and scientific explanation advanced by Bowers et al. for studies of vision using deep neural networks is also observed in several other domains of behavior research, and is in fact unavoidable when fitting large models such as deep nets and other supervised learners, with weak theoretical commitments, to restricted samples of highly stochastic behavioral phenomena.

Quality of concept maps is affected by map building strategies.

Concept-maps are widely used to assess students' conceptual understanding in different subject areas. Conventionally, it is mostly built maps which are assessed. In this study, we explore if "concept-mapping" could be used as a case for constructive alignment of both the process and the outcome of learning. Specifically, we have studied how a simple measure of looking at the order in which concept map elements-concepts (cards, C), links (arrows, A), and linking phrases (phrases, P)- are placed on the working space reveals information about the quality of the final generated map. We report findings from analysis of the concept-mapping process in two separate groups of university students (N = 38 (18 + 20)) who were asked, individually, to build concept maps related to two separate concepts in biology and chemistry respectively. We found that, across both groups, students consistently followed the same order of element placement that they began with and found significant differences in the quality of eventual maps resulting from students' map building strategies. Our results suggest that the quality of students' concept maps depends considerably on the strategy used to build it, and point to the supplementary role that the physical working space of the concept-mapping exercise offers to students' own working memory as a possible explanation for these quality differences.

Recombinant Enaction: Manipulatives Generate New Procedures in the Imagination, by Extending and Recombining Action Spaces.

Manipulation of physical models such as tangrams and tiles is a popular approach to teaching early mathematics concepts. This pedagogical approach is extended by new computational media, where mathematical entities such as equations and vectors can be virtually manipulated. The cognitive and neural mechanisms supporting such manipulation-based learning-particularly how actions generate new internal structures that support problem-solving-are not understood. We develop a model of the way manipulations generate internal traces embedding actions, and how these action-traces recombine during problem-solving. This model is based on a study of two groups of sixth-grade students solving area problems. Before problem-solving, one group manipulated a tangram, the other group answered a descriptive test. Eye-movement trajectories during problem-solving were different between the groups. A second study showed that this difference required the tangram's geometrical structure, just manipulation was not enough. We propose a theoretical model accounting for these results, and discuss its implications.

Intertemporal impulsivity can also arise from persistent failure of long-term plans.

We suggest that steep intertemporal discounting in individuals of low socioeconomic status (SES) may arise as a rational metacognitive adaptation to experiencing planning and control failures in long-term plans. Low SES individuals' plans fail more frequently because they operate close to budgetary boundaries, in turn because they consistently operate with limited budgets of money, status, trust, or other forms of social utility.

Attention Modulates Spatial Precision in Multiple-Object Tracking.

We present a computational model of multiple-object tracking that makes trial-level predictions about the allocation of visual attention and the effect of this allocation on observers' ability to track multiple objects simultaneously. This model follows the intuition that increased attention to a location increases the spatial resolution of its internal representation. Using a combination of empirical and computational experiments, we demonstrate the existence of a tight coupling between cognitive and perceptual resources in this task: Low-level tracking of objects generates bottom-up predictions of error likelihood, and high-level attention allocation selectively reduces error probabilities in attended locations while increasing it at non-attended locations. Whereas earlier models of multiple-object tracking have predicted the big picture relationship between stimulus complexity and response accuracy, our approach makes accurate predictions of both the macro-scale effect of target number and velocity on tracking difficulty and micro-scale variations in difficulty across individual trials and targets arising from the idiosyncratic within-trial interactions of targets and distractors.

Learning What to Want: Context-Sensitive Preference Learning.

We have developed a method for learning relative preferences from histories of choices made, without requiring an intermediate utility computation. Our method infers preferences that are rational in a psychological sense, where agent choices result from Bayesian inference of what to do from observable inputs. We further characterize conditions on choice histories wherein it is appropriate for modelers to describe relative preferences using ordinal utilities, and illustrate the importance of the influence of choice history by explaining all major categories of context effects using them. Our proposal clarifies the relationship between economic and psychological definitions of rationality and rationalizes several behaviors heretofore judged irrational by behavioral economists.

Modeling the repertoire of true tumor-specific MHC I epitopes in a human tumor.

DNA replication has a finite measurable error rate, net of repair, in all cells. Clonal proliferation of cancer cells leads therefore to accumulation of random mutations. A proportion of these mutational events can create new immunogenic epitopes that, if processed and presented by an MHC allele, may be recognized by the adaptive immune system. Here, we use probability theory to analyze the mutational and epitope composition of a tumor mass in successive division cycles and create a double Pölya model for calculating the number of truly tumor-specific MHC I epitopes in a human tumor. We deduce that depending upon tumor size, the degree of genomic instability and the degree of death within a tumor, human tumors have several tens to low hundreds of new, truly tumor-specific epitopes. Parenthetically, cancer stem cells, due to the asymmetry in their proliferative properties, shall harbor significantly fewer mutations, and therefore significantly fewer immunogenic epitopes. As the overwhelming majority of the mutations in cancer cells are unrelated to malignancy, the mutation-generated epitopes shall be specific for each individual tumor, and constitute the antigenic fingerprint of each tumor. These calculations highlight the benefits for personalization of immunotherapy of human cancer, and in view of the substantial pre-existing antigenic repertoire of tumors, emphasize the enormous potential of therapies that modulate the anti-cancer immune response by liberating it from inhibitory influences.