Continual learning enables incremental learning of new tasks without forgetting those previously learned, resulting in positive knowledge transfer that can enhance performance on both new and old tasks. However, continual learning poses new challenges for interpretability, as the rationale behind model predictions may change over time, leading to interpretability concept drift. We address this problem by proposing Interpretable Class-InCremental LEarning (ICICLE), an exemplar-free approach that adopts a prototypical part-based approach. It consists of three crucial novelties: interpretability regularization that distills previously learned concepts while preserving user-friendly positive reasoning; proximity-based prototype initialization strategy dedicated to the fine-grained setting; and task-recency bias compensation devoted to prototypical parts. Our experimental results demonstrate that ICICLE reduces the interpretability concept drift and outperforms the existing exemplar-free methods of common class-incremental learning when applied to concept-based models.

A film premiering next month will shine a light on a secret U.S. group which UFO researchers claim is helping to cover up the discovery of alien spacecraft. The film, God versus Aliens, has interviews with two experts about the'Collins Elite', a supposed secretive group within the U.S. military which has helped to cover up alien abductions and crashed spacecraft since the 1950s. British director of the film Mark Christopher Lee said: 'A lot of people know about Majestic 12, a supposed committee of military leaders and politicians interested in UFOs: it's out there in pop culture, along with Area 51. 'But two of my interviewees believe that this is a smokescreen, and the real organization is the Collins Elite, based in the Wright Patterson Air Base [in Ohio]. 'These people are said to work in a private organization on behalf of the government, because there are no freedom of information requests (FOIA) to private companies.' The Wright Patterson Air Base was home to the Project Blue Book investigation into UFO reports which began in 1947 - and there have been previous rumors of a secret'UFO room' at the base.

"AI for Good is about building AI in the right way and using it for social good. We've learned there are good business reasons for building this technology safely – if you want people to adopt it and use it, they need to be able to trust it. We've done a lot of work around helping domestic abuse victims in South Africa, with a chatbot called rAInbow. It was designed to help people understand their legal rights. It can be quite overwhelming to take that first step to getting help and trusted information if you don't know where to begin. I think it's important to acknowledge the risks this technology brings, but there are also tremendous positive opportunities. I spend a lot of my time building AI that helps improve the justice system and helps people understand their legal rights. With this technology, we can produce legal drafts in minutes that used to take days. Courts can function better and faster, so people can get their hearing dates and we can make the system more efficient. What excites me is that the new generation of technologists don't have to have my background. I went to geek school after geek school, but the newest programming language is human language. This means we can bring in people from many different backgrounds to build it. If we do this right, we will be opening up the profile of people who work in technology and AI." Fanning the flames of a "culture war" might drive ratings, generate clicks and provide politicians with election fodder, but the idea there are ever-deepening divides in British social attitudes is misleading.

Three Ukrainian drones were downed over Moscow early Sunday, Russia's defense ministry said, in an attack that briefly shut an international airport. While one of the drones was shot down on the city's outskirts, two others were "suppressed by electronic warfare" and smashed into an office complex. Moscow and its environs, lying about 500 kilometers from the Ukrainian border, had been rarely targeted during the conflict in Ukraine until several drone attacks this year.

Plastic has imbibed itself as an indispensable part of our day to day activities, becoming a source of problems due to its non-biodegradable nature and cheaper production prices. With these problems, comes the challenge of mitigating and responding to the aftereffects of disposal or the lack of proper disposal which leads to waste concentrating in locations and disturbing ecosystems for both plants and animals. As plastic debris levels continue to rise with the accumulation of waste in garbage patches in landfills and more hazardously in natural water bodies, swift action is necessary to plug or cease this flow. While manual sorting operations and detection can offer a solution, they can be augmented using highly advanced computer imagery linked with robotic appendages for removing wastes. The primary application of focus in this report are the much-discussed Computer Vision and Open Vision which have gained novelty for their light dependence on internet and ability to relay information in remote areas. These applications can be applied to the creation of edge-based mobility devices that can as a counter to the growing problem of plastic debris in oceans and rivers, demanding little connectivity and still offering the same results with reasonably timed maintenance. The principal findings of this project cover the various methods that were tested and deployed to detect waste in images, as well as comparing them against different waste types. The project has been able to produce workable models that can perform on time detection of sampled images using an augmented CNN approach. Latter portions of the project have also achieved a better interpretation of the necessary preprocessing steps required to arrive at the best accuracies, including the best hardware for expanding waste detection studies to larger environments.

IceCube DeepCore is an extension of the IceCube Neutrino Observatory designed to measure GeV scale atmospheric neutrino interactions for the purpose of neutrino oscillation studies. Distinguishing muon neutrinos from other flavors and reconstructing inelasticity are especially difficult tasks at GeV scale energies in IceCube DeepCore due to sparse instrumentation. Convolutional neural networks (CNNs) have been found to have better success at neutrino event reconstruction than conventional likelihood-based methods. In this contribution, we present a new CNN model that exploits time and depth translational symmetry in IceCube DeepCore data and present the model's performance, specifically for flavor identification and inelasticity reconstruction.

Machine learning practitioners often fine-tune generative pre-trained models like GPT-3 to improve model performance at specific tasks. Previous works, however, suggest that fine-tuned machine learning models memorize and emit sensitive information from the original fine-tuning dataset. Companies such as OpenAI offer fine-tuning services for their models, but no prior work has conducted a memorization attack on any closed-source models. In this work, we simulate a privacy attack on GPT-3 using OpenAI's fine-tuning API. Our objective is to determine if personally identifiable information (PII) can be extracted from this model. We (1) explore the use of naive prompting methods on a GPT-3 fine-tuned classification model, and (2) we design a practical word generation task called Autocomplete to investigate the extent of PII memorization in fine-tuned GPT-3 within a real-world context. Our findings reveal that fine-tuning GPT3 for both tasks led to the model memorizing and disclosing critical personally identifiable information (PII) obtained from the underlying fine-tuning dataset. To encourage further research, we have made our codes and datasets publicly available on GitHub at: https://github.com/albertsun1/gpt3-pii-attacks

Conformal prediction (CP) is a framework to quantify uncertainty of machine learning classifiers including deep neural networks. Given a testing example and a trained classifier, CP produces a prediction set of candidate labels with a user-specified coverage (i.e., true class label is contained with high probability). Almost all the existing work on CP assumes clean testing data and there is not much known about the robustness of CP algorithms w.r.t natural/adversarial perturbations to testing examples. This paper studies the problem of probabilistically robust conformal prediction (PRCP) which ensures robustness to most perturbations around clean input examples. PRCP generalizes the standard CP (cannot handle perturbations) and adversarially robust CP (ensures robustness w.r.t worst-case perturbations) to achieve better trade-offs between nominal performance and robustness. We propose a novel adaptive PRCP (aPRCP) algorithm to achieve probabilistically robust coverage. The key idea behind aPRCP is to determine two parallel thresholds, one for data samples and another one for the perturbations on data (aka "quantile-of-quantile" design). We provide theoretical analysis to show that aPRCP algorithm achieves robust coverage. Our experiments on CIFAR-10, CIFAR-100, and ImageNet datasets using deep neural networks demonstrate that aPRCP achieves better trade-offs than state-of-the-art CP and adversarially robust CP algorithms.

Concerns have been raised that they could be utilized to produce fake content with a deceptive intention, although evidence thus far remains anecdotal. This paper presents a case study about a Twitter botnet that appears to employ ChatGPT to generate human-like content. Through heuristics, we identify 1,140 accounts and validate them via manual annotation. These accounts form a dense cluster of fake personas that exhibit similar behaviors, including posting machine-generated content and stolen images, and engage with each other through replies and retweets. ChatGPT-generated content promotes suspicious websites and spreads harmful comments. While the accounts in the AI botnet can be detected through their coordination patterns, current state-of-the-art LLM content classifiers fail to discriminate between them and human accounts in the wild. These findings highlight the threats posed by AI-enabled social bots.

Adversarial examples threaten the integrity of machine learning systems with alarming success rates even under constrained black-box conditions. Stateful defenses have emerged as an effective countermeasure, detecting potential attacks by maintaining a buffer of recent queries and detecting new queries that are too similar. However, these defenses fundamentally pose a trade-off between attack detection and false positive rates, and this trade-off is typically optimized by hand-picking feature extractors and similarity thresholds that empirically work well. There is little current understanding as to the formal limits of this trade-off and the exact properties of the feature extractors/underlying problem domain that influence it. This work aims to address this gap by offering a theoretical characterization of the trade-off between detection and false positive rates for stateful defenses. We provide upper bounds for detection rates of a general class of feature extractors and analyze the impact of this trade-off on the convergence of black-box attacks. We then support our theoretical findings with empirical evaluations across multiple datasets and stateful defenses.