Machine learning can provide predictions with disparate outcomes, in which subgroups of the population (e.g., defined by age, gender, or other sensitive attributes) are systematically disadvantaged. In order to comply with upcoming legislation, practitioners need to locate such disparate outcomes. However, previous literature typically detects disparities through statistical procedures for when the sensitive attribute is specified a priori. This limits applicability in real-world settings where datasets are high dimensional and, on top of that, sensitive attributes may be unknown. As a remedy, we propose a data-driven framework called Automatic Location of Disparities (ALD) which aims at locating disparities in machine learning. ALD meets several demands from industry: ALD (1) is applicable to arbitrary machine learning classifiers; (2) operates on different definitions of disparities (e.g., statistical parity or equalized odds); and (3) deals with both categorical and continuous predictors even if disparities arise from complex and multi-way interactions known as intersectionality (e. g., age above 60 and female). ALD produces interpretable audit reports as output. We demonstrate the effectiveness of ALD based on both synthetic and real-world datasets. As a result, we empower practitioners to effectively locate and mitigate disparities in machine learning algorithms, conduct algorithmic audits, and protect individuals from discrimination.

Our interdisciplinary study investigates how effectively U.S. laws confront the challenges posed by Generative AI to human values. Through an analysis of diverse hypothetical scenarios crafted during an expert workshop, we have identified notable gaps and uncertainties within the existing legal framework regarding the protection of fundamental values, such as privacy, autonomy, dignity, diversity, equity, and physical/mental well-being. Constitutional and civil rights, it appears, may not provide sufficient protection against AI-generated discriminatory outputs. Furthermore, even if we exclude the liability shield provided by Section 230, proving causation for defamation and product liability claims is a challenging endeavor due to the intricate and opaque nature of AI systems. To address the unique and unforeseeable threats posed by Generative AI, we advocate for legal frameworks that evolve to recognize new threats and provide proactive, auditable guidelines to industry stakeholders. Addressing these issues requires deep interdisciplinary collaborations to identify harms, values, and mitigation strategies.

"Once the machine thinking method had started, it would not take long to outstrip our feeble powers. At some stage therefore we should have to expect the machines to take control" Alan Turing 1951 [35] AGI [91] safety is of the utmost urgency, since corporations and research labs are racing to build AGI despite prominent AI researchers and business leaders warning that it may lead to human extinction [11]. While governments are drafting AI regulations, there's little indication that they will be sufficient to resist competitive pressures and prevent the creation of AGI. Median estimates on the forecasting platform Metaculus of the date of AGI's creation have plummeted over the past few years from many decades away to 2027 [25] or 2032 [24] depending on definitions, with superintelligence expected to follow a few years later [23]. Is Alan Turing correct that we now "have to expect the machines to take control"?

Deep neural network representations align well with brain activity in the ventral visual stream. However, the primate visual system has a distinct dorsal processing stream with different functional properties. To test if a model trained to perceive 3D scene geometry aligns better with neural responses in dorsal visual areas, we trained a self-supervised geometry-aware recurrent neural network (GRNN) to predict novel camera views using a 3D feature memory. We compared GRNN to self-supervised baseline models that have been shown to align well with ventral regions using the large-scale fMRI Natural Scenes Dataset (NSD). We found that while the baseline models accounted better for ventral brain regions, GRNN accounted for a greater proportion of variance in dorsal brain regions. Our findings demonstrate the potential for using task-relevant models to probe representational differences across visual streams.

Language change is influenced by many factors, but often starts from synchronic variation, where multiple linguistic patterns or forms coexist, or where different speech communities use language in increasingly different ways. Besides regional or economic reasons, communities may form and segregate based on political alignment. The latter, referred to as political polarization, is of growing societal concern across the world. Here we map and quantify linguistic divergence across the partisan left-right divide in the United States, using social media data. We develop a general methodology to delineate (social) media users by their political preference, based on which (potentially biased) news media accounts they do and do not follow on a given platform. Our data consists of 1.5M short posts by 10k users (about 20M words) from the social media platform Twitter (now "X"). Delineating this sample involved mining the platform for the lists of followers (n=422M) of 72 large news media accounts. We quantify divergence in topics of conversation and word frequencies, messaging sentiment, and lexical semantics of words and emoji. We find signs of linguistic divergence across all these aspects, especially in topics and themes of conversation, in line with previous research. While US American English remains largely intelligible within its large speech community, our findings point at areas where miscommunication may eventually arise given ongoing polarization and therefore potential linguistic divergence. Our methodology - combining data mining, lexicostatistics, machine learning, large language models and a systematic human annotation approach - is largely language and platform agnostic. In other words, while we focus here on US political divides and US English, the same approach is applicable to other countries, languages, and social media platforms.

Malware is one of the most dangerous and costly cyber threats to national security and a crucial factor in modern cyber-space. However, the adoption of machine learning (ML) based solutions against malware threats has been relatively slow. Shortcomings in the existing ML approaches are likely contributing to this problem. The majority of current ML approaches ignore real-world challenges such as the detection of novel malware. In addition, proposed ML approaches are often designed either for malware/benign-ware classification or malware family classification. Here we introduce and showcase preliminary capabilities of a new method that can perform precise identification of novel malware families, while also unifying the capability for malware/benign-ware classification and malware family classification into a single framework.

The rapid advancements in artificial intelligence (AI) have led to a growing trend of human-AI teaming (HAT) in various fields. As machines continue to evolve from mere automation to a state of autonomy, they are increasingly exhibiting unexpected behaviors and human-like cognitive/intelligent capabilities, including situation awareness (SA). This shift has the potential to enhance the performance of mixed human-AI teams over all-human teams, underscoring the need for a better understanding of the dynamic SA interactions between humans and machines. To this end, we provide a review of leading SA theoretical models and a new framework for SA in the HAT context based on the key features and processes of HAT. The Agent Teaming Situation Awareness (ATSA) framework unifies human and AI behavior, and involves bidirectional, and dynamic interaction. The framework is based on the individual and team SA models and elaborates on the cognitive mechanisms for modeling HAT. Similar perceptual cycles are adopted for the individual (including both human and AI) and the whole team, which is tailored to the unique requirements of the HAT context. ATSA emphasizes cohesive and effective HAT through structures and components, including teaming understanding, teaming control, and the world, as well as adhesive transactive part. We further propose several future research directions to expand on the distinctive contributions of ATSA and address the specific and pressing next steps.

Machine-learning models are known to be vulnerable to evasion attacks that perturb model inputs to induce misclassifications. In this work, we identify real-world scenarios where the true threat cannot be assessed accurately by existing attacks. Specifically, we find that conventional metrics measuring targeted and untargeted robustness do not appropriately reflect a model's ability to withstand attacks from one set of source classes to another set of target classes. To address the shortcomings of existing methods, we formally define a new metric, termed group-based robustness, that complements existing metrics and is better-suited for evaluating model performance in certain attack scenarios. We show empirically that group-based robustness allows us to distinguish between models' vulnerability against specific threat models in situations where traditional robustness metrics do not apply. Moreover, to measure group-based robustness efficiently and accurately, we 1) propose two loss functions and 2) identify three new attack strategies. We show empirically that with comparable success rates, finding evasive samples using our new loss functions saves computation by a factor as large as the number of targeted classes, and finding evasive samples using our new attack strategies saves time by up to 99\% compared to brute-force search methods. Finally, we propose a defense method that increases group-based robustness by up to 3.52$\times$.

Following four successful years in the SAE AutoDrive Challenge Series I, the University of Toronto is participating in the Series II competition to develop a Level 4 autonomous passenger vehicle capable of handling various urban driving scenarios by 2025. Accurate detection of traffic lights and correct identification of their states is essential for safe autonomous operation in cities. Herein, we describe our recently-redesigned traffic light perception system for autonomous vehicles like the University of Toronto's self-driving car, Artemis. Similar to most traffic light perception systems, we rely primarily on camera-based object detectors. We deploy the YOLOv5 detector for bounding box regression and traffic light classification across multiple cameras and fuse the observations. To improve robustness, we incorporate priors from high-definition semantic maps and perform state filtering using hidden Markov models. We demonstrate a multi-camera, real time-capable traffic light perception pipeline that handles complex situations including multiple visible intersections, traffic light variations, temporary occlusion, and flashing light states. To validate our system, we collected and annotated a varied dataset incorporating flashing states and a range of occlusion types. Our results show superior performance in challenging real-world scenarios compared to single-frame, single-camera object detection.

Data collected at Hurricane Ian (2022) quantifies the demands that small uncrewed aerial systems (UAS), or drones, place on the network communication infrastructure and identifies gaps in the field. Drones have been increasingly used since Hurricane Katrina (2005) for disaster response, however getting the data from the drone to the appropriate decision makers throughout incident command in a timely fashion has been problematic. These delays have persisted even as countries such as the USA have made significant investments in wireless infrastructure, rapidly deployable nodes, and an increase in commercial satellite solutions. Hurricane Ian serves as a case study of the mismatch between communications needs and capabilities. In the first four days of the response, nine drone teams flew 34 missions under the direction of the State of Florida FL-UAS1, generating 636GB of data. The teams had access to six different wireless communications networks but had to resort to physically transferring data to the nearest intact emergency operations center in order to make the data available to the relevant agencies. The analysis of the mismatch contributes a model of the drone data-to-decision workflow in a disaster and quantifies wireless network communication requirements throughout the workflow in five factors. Four of the factors-availability, bandwidth, burstiness, and spatial distribution-were previously identified from analyses of Hurricanes Harvey (2017) and Michael (2018). This work adds upload rate as a fifth attribute. The analysis is expected to improve drone design and edge computing schemes as well as inform wireless communication research and development.