Declan was so shocked he didn't say anything, and for the rest of the session he was privy to a real-time stream of ChatGPT analysis rippling across his therapist's screen. The session became even more surreal when Declan began echoing ChatGPT in his own responses, preempting his therapist. "I became the best patient ever," he says, "because ChatGPT would be like, 'Well, do you consider that your way of thinking might be a little too black and white?' And I would be like, 'Huh, you know, I think my way of thinking might be too black and white,' and [my therapist would] be like, 'Exactly.' I'm sure it was his dream session."

In 2017, a game design thinktank called Project Horseshoe gathered a group of developers together to define the concept of cosiness in video games. Games, of course, have had non-violent elements since the medium was invented. Early life simulators such as 1985's Little Computer People, a low-stakes game in which the player interacts with a man living his unremarkable life in a house, could fit the bill; then there was the proliferation of social farming simulations after 1996's chibi-adorable Harvest Moon. But the resulting report, Coziness in Games: An Exploration of Safety, Softness, and Satisfied Needs, is probably the first organised effort to define a then-emerging genre. Cosy games (cozy in US spelling) don't have high-risk scenarios: "There is no impending loss of threat," they wrote.

Although value-aligned language models (LMs) appear unbiased in explicit bias evaluations, they often exhibit stereotypes in implicit word association tasks, raising concerns about their fair usage. We investigate the mechanisms behind this discrepancy and find that alignment surprisingly amplifies implicit bias in model outputs. Specifically, we show that aligned LMs, unlike their unaligned counterparts, overlook racial concepts in early internal representations when the context is ambiguous. Not representing race likely fails to activate safety guardrails, leading to unintended biases. Inspired by this insight, we propose a new bias mitigation strategy that works by incentivizing the representation of racial concepts in the early model layers. In contrast to conventional mitigation methods of machine unlearning, our interventions find that steering the model to be more aware of racial concepts effectively mitigates implicit bias. Similar to race blindness in humans, ignoring racial nuances can inadvertently perpetuate subtle biases in LMs.

Distributed optimization is fundamental to modern machine learning applications like federated learning, but existing methods often struggle with ill-conditioned problems and face stability-versus-speed tradeoffs. We introduce fractional order distributed optimization (FrODO); a theoretically-grounded framework that incorporates fractional-order memory terms to enhance convergence properties in challenging optimization landscapes. Our approach achieves provable linear convergence for any strongly connected network. Through empirical validation, our results suggest that FrODO achieves up to 4 times faster convergence versus baselines on ill-conditioned problems and 2-3 times speedup in federated neural network training, while maintaining stability and theoretical guarantees.

To mitigate the risk of harmful outputs from large vision models (LVMs), we introduce the SafeSora dataset to promote research on aligning text-to-video generation with human values. This dataset encompasses human preferences in text-to-video generation tasks along two primary dimensions: helpfulness and harmlessness. To capture in-depth human preferences and facilitate structured reasoning by crowdworkers, we subdivide helpfulness into 4 sub-dimensions and harmlessness into 12 sub-categories, serving as the basis for pilot annotations. The SafeSora dataset includes 14,711 unique prompts, 57,333 unique videos generated by 4 distinct LVMs, and 51,691 pairs of preference annotations labeled by humans. We further demonstrate the utility of the SafeSora dataset through several applications, including training the text-video moderation model and aligning LVMs with human preference by fine-tuning a prompt augmentation module or the diffusion model. These applications highlight its potential as the foundation for text-to-video alignment research, such as human preference modeling and the development and validation of alignment algorithms.

Bistable images, also known as ambiguous or reversible images, present visual stimuli that can be seen in two distinct interpretations, though not simultaneously by the observer. In this study, we conduct the most extensive examination of vision-language models using bistable images to date. We manually gathered a dataset of 29 bistable images, along with their associated labels, and subjected them to 116 different manipulations in brightness, tint, and rotation. We evaluated twelve different models in both classification and generative tasks across six model architectures. Our findings reveal that, with the exception of models from the Idefics family and LLaVA1.5-13b, there is a pronounced preference for one interpretation over another among the models, and minimal variance under image manipulations, with few exceptions on image rotations. Additionally, we compared the model preferences with humans, noting that the models do not exhibit the same continuity biases as humans and often diverge from human initial interpretations. We also investigated the influence of variations in prompts and the use of synonymous labels, discovering that these factors significantly affect model interpretations more than image manipulations showing a higher influence of the language priors on bistable image interpretations compared to image-text training data. All code and data is open sourced.

AI Art Generator App. ✅ Fast ✅ Free ✅ Easy. Create amazing artworks using artificial intelligence.

If you spend too much time on your smartphone, scientists have a list of 10 solutions that can help you cut back on screen time. The small but effective changes can help curb smartphone addiction and mental health issues such as depression, say experts at McGill University in Canada. In experiments, people following the strategies reduced their screen time, felt less addicted to their phone and improved their sleep quality, the experts report. Among the 10 strategies are changing the phone display to'greyscale' so the display appears black and white, and disabling facial recognition as a method of unlocking the screen. A black and white screen makes smartphones'less gratifying' to look at compared to the bright colours offered by app icons such as TikTok and Instagram.

Phishing attacks are cyber-attacks through which criminals trick users into sending them money and sensitive information, or into installing malware on their computer, by sending them deceptive emails or messages. As these attacks have become increasingly widespread, developers have been trying to develop more advanced tools to detect them and protect potential victims. Researchers at Monash University and CSIRO's Data61 in Australia have recently developed a machine learning-based approach that could help users to identify phishing emails, so that they don't inadvertently install malware or send sensitive data to cyber-criminals. This model was introduced in a paper pre-published on arXiv and set to be presented at AsiaCCS 2022, a cyber-security conference. "We have identified a gap in current phishing research, namely realizing that existing literature focuses on rigorous'black and white' methods to classify whether something is a phishing email or not," Tingmin (Tina) Wu, one of the researchers who carried out the study, told TechXplore.