More Like this

1. Community-Aware Group Testing

   https://doi.org/10.1109/TIT.2023.3250119  

   Nikolopoulos, Pavlos; Srinivasavaradhan, Sundara Rajan; Guo, Tao; Fragouli, Christina; Diggavi, Suhas N. (July 2023, IEEE Transactions on Information Theory)

   Group testing is a technique that can reduce the number of tests needed to identify infected members in a population, by pooling together multiple diagnostic samples. Despite the variety and importance of prior results, traditional work on group testing has typically assumed independent infections. However, contagious diseases among humans, like SARS-CoV-2, have an important characteristic: infections are governed by community spread, and are therefore correlated. In this paper, we explore this observation and we argue that taking into account the community structure when testing can lead to significant savings in terms of the number of tests required to guarantee a given identification accuracy. To show that, we start with a simplistic (yet practical) infection model, where the entire population is organized in (possibly overlapping) communities and the infection probability of an individual depends on the communities (s)he participates in. Given this model, we compute new lower bounds on the number of tests for zero-error identification and design community-aware group testing algorithms that can be optimal under assumptions. Finally, we demonstrate significant benefits over traditional, community-agnostic group testing via simulations using both noiseless and noisy tests 

   more » « less
2. Testing for periodicity in functional time series

   https://doi.org/DOI: 10.1214/17-AOS1645  

   H; Kokoszka, P.; Nisol, G. (January 2018, Annals of statistics)

   We derive several tests for the presence of a periodic component in a time series of functions. We consider both the traditional setting in which the periodic functional signal is contaminated by functional white noise, and a more general setting of a weakly dependent contaminating process. Several forms of the periodic component are considered. Our tests are motivated by the likelihood principle and fall into two broad categories, which we term multivariate and fully functional. Generally, for the functional series that motivate this research, the fully functional tests exhibit a superior balance of size and power. Asymptotic null distributions of all tests are derived and their consistency is established. Their finite sample performance is examined and compared by numerical studies and application to pollution data. 

   more » « less
3. Spontaneous Learning of Face Identity in Expression-Trained Deep Nets

   https://doi.org/10.32470/CCN.2022.1116-0  

   Schwartz, Emily; O'Nell, Kathryn; Saxe, Rebecca; Anzellotti, Stefano (January 2022, Conference on Cognitive Computational Neuroscience)

   Recent neural evidence challenges the traditional view that face identity and facial expressions are processed by segregated neural pathways, showing that information about identity and expression are encoded within common brain regions. This article tests the hypothesis that integrated representations of identity and expression arise naturally within neural networks. Deep networks trained to recognize expression and deep networks trained to recognize identity spontaneously develop representations of identity and expression, respectively. These findings serve as a “proof-of-concept” that it is not necessary to discard task-irrelevant information for identity and expression recognition. 

   more » « less
4. Inference at Scale: Significance Testing for Large Search and Recommendation Experiments

   Ihemelandu, Ngozi; Ekstrand, Michael D. (July 2023, Proceedings of the 46th International ACM SIGIR Conference on Research and Development in Information Retrieval (SIGIR '23))

   A number of information retrieval studies have been done to assess which statistical techniques are appropriate for comparing systems. However, these studies are focused on TREC-style experiments, which typically have fewer than 100 topics. There is no similar line of work for large search and recommendation experiments; such studies typically have thousands of topics or users and much sparser relevance judgements, so it is not clear if recommendations for analyzing traditional TREC experiments apply to these settings. In this paper, we empirically study the behavior of significance tests with large search and recommendation evaluation data. Our results show that the Wilcoxon and Sign tests show significantly higher Type-1 error rates for large sample sizes than the bootstrap, randomization and t-tests, which were more consistent with the expected error rate. While the statistical tests displayed differences in their power for smaller sample sizes, they showed no difference in their power for large sample sizes. We recommend the sign and Wilcoxon tests should not be used to analyze large scale evaluation results. Our result demonstrate that with Top-\(N\) recommendation and large search evaluation data, most tests would have a 100% chance of finding statistically significant results. Therefore, the effect size should be used to determine practical or scientific significance. 

   more » « less
5. Finite-Sample Bounds for Two-Distribution Hypothesis Tests

   https://doi.org/10.1109/DSAA60987.2023.10302643  

   Hom, Cynthia; Yik, William; Montañez, George D. (October 2023, 2023 IEEE 10th International Conference on Data Science and Advanced Analytics (DSAA))

   With the rapid growth of large language models, big data, and malicious online attacks, it has become increasingly important to have tools for anomaly detection that can distinguish machine from human, fair from unfair, and dangerous from safe. Prior work has shown that two-distribution (specified complexity) hypothesis tests are useful tools for such tasks, aiding in detecting bias in datasets and providing artificial agents with the ability to recognize artifacts that are likely to have been designed by humans and pose a threat. However, existing work on two-distribution hypothesis tests requires exact values for the specification function, which can often be costly or impossible to compute. In this work, we prove novel finite-sample bounds that allow for two-distribution hypothesis tests with only estimates of required quantities, such as specification function values. Significantly, the resulting bounds do not require knowledge of the true distribution, distinguishing them from traditional p-values. We apply our bounds to detect student cheating on multiple-choice tests, as an example where the exact specification function is unknown. We additionally apply our results to detect representational bias in machine-learning datasets and provide artificial agents with intention perception, showing that our results are consistent with prior work despite only requiring a finite sample of the space. Finally, we discuss additional applications and provide guidance for those applying these bounds to their own work. 

   more » « less