相关论文: The Moon Illusion explained by the Projective Cons…

In this paper, we investigate the possibility that a geometric model of dark energy that assume \textquotedblleft lower dimensions\textquotedblright, could generates a repulsive pressure which cause the current accelerated expansion of the…

Large language models are successful in answering factoid questions but are also prone to hallucination. We investigate the phenomenon of LLMs possessing correct answer knowledge yet still hallucinating from the perspective of inference…

Humans are able to conceive physical reality by jointly learning different facets thereof. To every pair of notions related to a perceived reality may correspond a mutual relation, which is a notion on its own, but one-level higher. Thus,…

Motivated by the observed cosmic matter distribution, we present the following conjecture: due to the formation of voids and opaque structures, the average matter density on the path of the light from the well-observed objects changes from…

A new prior is proposed for learning representations of high-level concepts of the kind we manipulate with language. This prior can be combined with other priors in order to help disentangling abstract factors from each other. It is…

According to the giant impact theory, the Moon formed through accreting the debris disk produced by a collision between Theia and the proto-Earth, and the predicted lunar orbital inclination relative to the Earth's equatorial plane is about…

Modern language models can imitate complex patterns through few-shot learning, enabling them to complete challenging tasks without fine-tuning. However, imitation can also lead models to reproduce inaccuracies or harmful content if present…

In functional analysis there are several reasonable approaches to the notion of a projective module. We show that a certain general-categorical framework contains, as particular cases, all known versions. In this scheme, the notion of a…

Attention mechanisms are ubiquitous components in neural architectures applied to natural language processing. In addition to yielding gains in predictive accuracy, attention weights are often claimed to confer interpretability, purportedly…

Input space reconstruction is an attractive representation learning paradigm. Despite interpretability of the reconstruction and generation, we identify a misalignment between learning by reconstruction, and learning for perception. We show…

What makes an image appear realistic? In this work, we are answering this question from a data-driven perspective by learning the perception of visual realism directly from large amounts of data. In particular, we train a Convolutional…

Human planning is efficient--it frugally deploys limited cognitive resources to accomplish difficult tasks--and flexible--adapting to novel problems and environments. Computational approaches suggest that people construct simplified mental…

We propose to use transformation optics to generate a general illusion such that an arbitrary object appears to be like some other object of our choice. This is achieved by using a remote device that transforms the scattered light outside a…

Because we are highly motivated to be understood, we created public external representations -- mime, language, art -- to externalise our inner states. We argue that such external representations are a pre-condition for access…

Random Projection is a foundational research topic that connects a bunch of machine learning algorithms under a similar mathematical basis. It is used to reduce the dimensionality of the dataset by projecting the data points efficiently to…

Earthshine is the dominant source of natural illumination on the surface of the Moon during lunar night, and at locations within permanently shadowed regions that never receive direct sunlight. As such, earthshine may enable the exploration…

Latent visual reasoning aims to mimic human's imagination process by meditating through hidden states of Multimodal Large Language Models. While recognized as a promising paradigm for visual reasoning, the underlying mechanisms driving its…

Many techniques in machine learning attempt explicitly or implicitly to infer a low-dimensional manifold structure of an underlying physical phenomenon from measurements without an explicit model of the phenomenon or the measurement…

The Platonic Representation Hypothesis claims that recent foundation models are converging to a shared representation space as a function of their downstream task performance, irrespective of the objectives and data modalities used to train…

When measuring the mass profile of any given cosmological structure through internal kinematics, the distant background density is always ignored. This trick is often refereed to as the "Jeans Swindle". Without this trick a divergent term…