Transformer-based object detectors have shown competitive performance recently. Compared with convolutional neural networks limited by the relatively small receptive fields, the advantage of transformer for visual tasks is the capacity to perceive long-range dependencies among all image patches, while the deficiency is that the local fine-grained information is not fully excavated. In this paper, we introduce the Coarse-grained and Fine-grained crossing representations to build an efficient Detection Transformer (CFDT). Specifically, we propose a local-global cross fusion module to establish the connection between local fine-grained features and global coarse-grained features. Besides, we propose a coarse-fine aware neck which enables detection tokens to interact with both coarse-grained and fine-grained features. Furthermore, an efficient feature integration module is presented for fusing multi-scale representations from different stages. Experimental results on the COCO dataset demonstrate the effectiveness of the proposed method. For instance, our CFDT achieves 48.1 AP with 173G FLOPs, which possesses higher accuracy and less computation compared with the state-of-the-art transformer-based detector ViDT.

Tensor multiplication with learned weight matrices is the fundamental building block in deep learning models. These matrices can often be sparsified, decomposed, quantized, or subjected to random parameter sharing without losing accuracy, suggesting the possibility of more efficient transforms. Although many variants of weight matrices exist, unstructured ones are incompatible with modern hardware, slowing inference and training. On the other hand, structured variants often limit expressivity or fail to deliver the promised latency benefits. We present Sketch Structured Transform (SS1), an expressive and GPU-friendly operator that accelerates inference.

What representation do deep neural networks learn? How similar are images to each other for neural networks? Despite the overwhelming success of deep learning methods key questions about their internal workings still remain largely unanswered, due to their internal high dimensionality and complexity. To address this, one approach is to measure the similarity of activation responses to various inputs.Representational Similarity Matrices (RSMs) distill this similarity into scalar values for each input pair.These matrices encapsulate the entire similarity structure of a system, indicating which input lead to similar responses.While the similarity between images is ambiguous, we argue that the spatial location of semantic objects does neither influence human perception nor deep learning classifiers. Thus this should be reflected in the definition of similarity between image responses for computer vision systems. Revisiting the established similarity calculations for RSMs we expose their sensitivity to spatial alignment. In this paper we propose to solve this through, which are invariant to spatial permutation.

Delivering precise point and distributional forecasts across a spectrum of prediction horizons represents a significant and enduring challenge in the application of time-series forecasting within various industries.Prior research on developing deep learning models for time-series forecasting has often concentrated on isolated aspects, such as long-term point forecasting or short-term probabilistic estimations. This narrow focus may result in skewed methodological choices and hinder the adaptability of these models to uncharted scenarios.While there is a rising trend in developing universal forecasting models, a thorough understanding of their advantages and drawbacks, especially regarding essential forecasting needs like point and distributional forecasts across short and long horizons, is still lacking.In this paper, we present ProbTS, a benchmark tool designed as a unified platform to evaluate these fundamental forecasting needs and to conduct a rigorous comparative analysis of numerous cutting-edge studies from recent years.We dissect the distinctive data characteristics arising from disparate forecasting requirements and elucidate how these characteristics can skew methodological preferences in typical research trajectories, which often fail to fully accommodate essential forecasting needs.Building on this, we examine the latest models for universal time-series forecasting and discover that our analyses of methodological strengths and weaknesses are also applicable to these universal models.Finally, we outline the limitations inherent in current research and underscore several avenues for future exploration.

PCWorld reports that AI is evolving beyond simple chatbots to become autonomous agents that directly control PC functions and applications. Major tech companies are developing agentic AI systems, including Anthropic's Claude tools, OpenAI's upcoming superapp, and Google's Gemini Mac app with desktop intelligence features. This shift toward AI agents managing tasks like software development and data analysis represents a fundamental change in how users will interact with their computers. Remember when ChatGPT was just an AI chatbox that sat on your desktop? That was, like, so December.

OpenAI is developing a unified desktop superapp that will integrate ChatGPT, Codex, and Atlas into a single application, according to PCWorld's coverage of The Wall Street Journal report. This consolidation aims to reduce service fragmentation and improve overall quality for users accessing OpenAI's various AI tools. The superapp represents a significant shift toward streamlined AI services, potentially making OpenAI's offerings more accessible and efficient for desktop users. It seems you'll soon be able to access most of OpenAI's services in one place on your computer.

Tech companies believe in intellectual property, but not yours. In April 2024, Eric Schmidt, the former Google CEO and a current AI evangelist, gave a closed-door lecture to a group of Stanford students. If these young people hoped to be Silicon Valley entrepreneurs, Schmidt explained, then they should be prepared to breach some ethical boundaries. Yet Schmidt told the students to go ahead and download whatever they need to build an accurate "test" version of their AI product. If the product takes off, "then you hire a whole bunch of lawyers to go clean the mess up," he said.

At Palantir's Developer Conference, AI Is Built to Win Wars As business soars, Palantir is doubling down on a vision of AI built for battlefield advantage--and attracting customers who agree. The defense contractors, military officers, and corporate executives in attendance are unprepared for the weather; they'd assumed the previous day's mid-70s temperatures would hold. A cold rain turns to steady snowfall, and Palantir passes out heavy blankets. As people move between open-air pavilions, it looks like they were pulled from shipwrecks. To this self-selecting crowd, Palantir is delivering on its promises.

Modern machine learning heavily depends on the effectiveness of optimization techniques. While deep learning models have achieved remarkable empirical results in training, their theoretical underpinnings remain somewhat elusive. Ensuring the convergence of optimization methods requires imposing specific structures on the objective function which often do not hold in practice. One prominent example is the widely recognized Polyak-Lojasiewicz (PL) inequality, which has garnered considerable attention in recent years. However, validating such assumptions for deep neural networks entails substantial and often impractical levels of over-parametrization. In order to address this limitation, we propose a novel class of functions that can characterize the loss landscape of modern deep models without requiring extensive over-parametrization and can also include saddle points. Crucially, we prove that gradient-based optimizers possess theoretical guarantees of convergence under this assumption.

At the core of 3D PTV is the dual-frame fluid motion estimation algorithm, which tracks particles across two consecutive frames. Recently, deep learning-based methods have achieved impressive accuracy in dual-frame fluid motion estimation; however, they heavily depend on large volumes of labeled data. In this paper, we introduce a new method that is completely self-supervised and notably outperforms its fully-supervised counterparts while requiring only 1\% of the training samples (without labels) used by previous methods. Our method features a novel zero-divergence loss that is specific to the domain of turbulent flow. Inspired by the success of splat operation in high-dimensional filtering and random fields, we propose a splat-based implementation for this loss which is both efficient and effective. The self-supervised nature of our method naturally supports test-time optimization, leading to the development of a tailored Dynamic Velocimetry Enhancer (DVE) module. We demonstrate that strong cross-domain robustness is achieved through test-time optimization on unseen leave-one-out synthetic domains and real physical/biological domains.