In the early 1900s, beavers had almost completely disappeared from the United States due to hunting and trapping. Breakthroughs, discoveries, and DIY tips sent every weekday. Beavers might rival even the most hardworking corporate employee in productivity and hustle, but they're not quite cut out for business travel--especially the airborne kind. Nevertheless, in 1948, 76 industrious beavers were subjected to a once-in-a-lifetime "work trip" to Idaho's remote Chamberlain Basin--via parachute. The event, which was captured in a now-viral video, has become celebrated as a quirky example of human ingenuity and environmental stewardship. After all, who can resist a flying beaver?

The impressive capability of modern text-to-image models to generate realistic visuals has come with a serious drawback: they can be misused to create harmful, deceptive or unlawful content. This has accelerated the push for machine unlearning. This new field seeks to selectively remove specific knowledge from a model's training data without causing a drop in its overall performance. However, it turns out that actually forgetting a given concept is an extremely difficult task. Models exposed to attacks using adversarial prompts show the ability to generate so-called unlearned concepts, which can be not only harmful but also illegal. In this paper, we present considerations regarding the ability of models to forget and recall knowledge, introducing the Memory Self-Regeneration task. Furthermore, we present MemoRa strategy, which we consider to be a regenerative approach supporting the effective recovery of previously lost knowledge. Moreover, we propose that robustness in knowledge retrieval is a crucial yet underexplored evaluation measure for developing more robust and effective unlearning techniques. Finally, we demonstrate that forgetting occurs in two distinct ways: short-term, where concepts can be quickly recalled, and long-term, where recovery is more challenging. Code is available at https://gmum.github.io/MemoRa/.

With the surge and widespread application of image generation models, data privacy and content safety have become major concerns and attracted great attention from users, service providers, and policymakers. Machine unlearning (MU) is recognized as a cost-effective and promising means to address these challenges. Despite some advancements, image generation model unlearning (IGMU) still faces remarkable gaps in practice, e.g., unclear task discrimination and unlearning guidelines, lack of an effective evaluation framework, and unreliable evaluation metrics. These can hinder the understanding of unlearning mechanisms and the design of practical unlearning algorithms. We perform exhaustive assessments over existing state-of-the-art unlearning algorithms and evaluation standards, and discover several critical flaws and challenges in IGMU tasks. Driven by these limitations, we make several core contributions, to facilitate the comprehensive understanding, standardized categorization, and reliable evaluation of IGMU. Specifically, (1) We design CatIGMU, a novel hierarchical task categorization framework. It provides detailed implementation guidance for IGMU, assisting in the design of unlearning algorithms and the construction of testbeds. (2) We introduce EvalIGMU, a comprehensive evaluation framework. It includes reliable quantitative metrics across five critical aspects. (3) We construct DataIGM, a high-quality unlearning dataset, which can be used for extensive evaluations of IGMU, training content detectors for judgment, and benchmarking the state-of-the-art unlearning algorithms. With EvalIGMU and DataIGM, we discover that most existing IGMU algorithms cannot handle the unlearning well across different evaluation dimensions, especially for preservation and robustness. Code and models are available at https://github.com/ryliu68/IGMU.

Concept erasing has recently emerged as an effective paradigm to prevent text-to-image diffusion models from generating visually undesirable or even harmful content. However, current removal methods heavily rely on manually crafted text prompts, making it challenging to achieve a high erasure (efficacy) while minimizing the impact on other benign concepts (usability). In this paper, we attribute the limitations to the inherent gap between the text and image modalities, which makes it hard to transfer the intricately entangled concept knowledge from text prompts to the image generation process. To address this, we propose a novel solution by directly integrating visual supervision into the erasure process, introducing the first text-image Collaborative Concept Erasing (Co-Erasing) framework. Specifically, Co-Erasing describes the concept jointly by text prompts and the corresponding undesirable images induced by the prompts, and then reduces the generating probability of the target concept through negative guidance. This approach effectively bypasses the knowledge gap between text and image, significantly enhancing erasure efficacy. Additionally, we design a text-guided image concept refinement strategy that directs the model to focus on visual features most relevant to the specified text concept, minimizing disruption to other benign concepts. Finally, comprehensive experiments suggest that Co-Erasing outperforms state-of-the-art erasure approaches significantly with a better trade-off between efficacy and usability. Codes are available at https://github.com/Ferry-Li/Co-Erasing.

Transformer-based models have dominated natural language processing and other areas in the last few years due to their superior (zero-shot) performance on benchmark datasets. However, these models are poorly understood due to their complexity and size. While probing-based methods are widely used to understand specific properties, the structures of the representation space are not systematically characterized; consequently, it is unclear how such models generalize and overgeneralize to new inputs beyond datasets. In this paper, based on a new gradient descent optimization method, we are able to explore the embedding space of a commonly used vision-language model. Using the Imagenette dataset, we show that while the model achieves over 99\% zero-shot classification performance, it fails systematic evaluations completely. Using a linear approximation, we provide a framework to explain the striking differences. We have also obtained similar results using a different model to support that our results are applicable to other transformer models with continuous inputs. We also propose a robust way to detect the modified images.

The Call of Duty devs are always trying to stay one step ahead of cheaters to protect the experience for all of us regular non-jerky players. Their latest move to prevent cheating may just be the funniest one yet. The devs have announced an appropriately-named feature called Splat. When a cheater deploys, the system occasionally disables their parachute, sending them careening to the ground until they, well, go splat. This was designed to call as much attention to the cheater as possible, with devs saying it'll be "immediately obvious" who's trying to game the system, as someone quickly descending from the sky is hard to miss.

The United States has accused Russian fighter jets of flying dangerously close to several of its drones over Syria, setting off flares and forcing the MQ-9 Reapers to take evasive action. US Air Forces Central released a video of Wednesday's encounter, showing a Russian SU-35 fighter closing in on the drone. Footage showed the Russian pilot positioning his aircraft in front of the Reaper and turning on the afterburner, dramatically increasing speed and air pressure and making it harder to operate the drone, the air force said in comments accompanying the video. So-called parachute flares were also released. "The Russian SU-35 fighter aircraft employed parachute flares in the flight path of US MQ-9 aircraft," the air force said.

A surge of advances in language models (LMs) has led to significant interest in using LMs to build co-writing systems, in which humans and LMs interactively contribute to a shared writing artifact. However, there is a lack of studies assessing co-writing systems in interactive settings. We propose a human-centered evaluation framework, Parachute, for interactive co-writing systems. Parachute showcases an integrative view of interaction evaluation, where each evaluation aspect consists of categorized practical metrics. Furthermore, we present Parachute with a use case to demonstrate how to evaluate and compare co-writing systems using Parachute.

A drone that apparently flew all the way from the Ukrainian war zone has crashed on the outskirts of the Croatian capital, Zagreb, triggering a loud blast but causing no injuries, according to Croatian authorities. A statement issued on Friday after Croatia's National Security Council meeting said the "pilotless military aircraft" entered Croatian airspace overnight from neighbouring Hungary at a speed of 700 kilometres per hour (430mph) and an altitude of 1,300 metres (4,300 feet). The council said that an official criminal investigation will be launched and that NATO will be informed about the incident. The crash means that the large drone flew at least 560km (350 miles) apparently undetected by air defences in Croatia and Hungary, both members of the Western military alliance. Military experts of The War Zone online magazine said that the aircraft is likely a Soviet-era Tu-141 Strizh reconnaissance drone that must have severely malfunctioned and crossed over the entirety of Hungary and into Croatia from Ukraine.

Kentucky-based Aviation Safety Resources is developing ballistic parachutes for use in aircraft ranging from 60 lbs to 12,000 lbs. The Air Force's Agility Prime program awarded a phase I small business technology transfer (STTR) research contract to Jump Aero and Caltech to create an electronic parachute powered by machine learning that would allow the pilot to recalibrate the flight controller in midair in the event of damage, the company announced on April 7. "The electronic parachute is the name for the concept of implementing an adaptive/machine-learned control routine that would be impractical to certify for the traditional controller for use only in an emergency recovery mode -- something that would be switched on by the pilot if there is reason to believe that the baseline flight controller is not properly controlling the aircraft (if, for example, the aircraft has been damaged in midair)," Carl Dietrich, founder and president of Jump Aero Incorporated, told Avionics International. This technology was previously difficult to certify because of the need for deterministic proof of safety within these complex systems. The research was sparked when the Federal Aviation Administration certified an autonomous landing function for use in emergency situations which created a path for the possible certification of electronic parachute technology, according to Jump Aero. The machine-learned neural network can be trained with non-linear behaviors that occur in an aircraft in the presence of substantial failures such those generated by a bird strike, Dietrich said.