The creation of new approaches to the design and configuration of smart buildings relies heavily on AI tools and Machine Learning (ML) algorithms, particularly optimization techniques. The widespread use of electronic...
详细信息
Recent advancements in Visual Question Answering (VQA) have been driven by the integration of complex attention mechanisms. This work introduces a novel approach aimed at enhancing multi-modal representations through ...
详细信息
The transformer architecture has prevailed in various deep learning settings due to its exceptional capabilities to select and compose structural information. Motivated by these capabilities, Sanford et al. (2023) pro...
详细信息
The transformer architecture has prevailed in various deep learning settings due to its exceptional capabilities to select and compose structural information. Motivated by these capabilities, Sanford et al. (2023) proposed the sparse token selection task, in which transformers excel while fully-connected networks (FCNs) fail in the worst case. Building upon that, we strengthen the FCN lower bound to an average-case setting and establish an algorithmic separation of transformers over FCNs. Specifically, a one-layer transformer trained with gradient descent provably learns the sparse token selection task and, surprisingly, exhibits strong out-of-distribution length generalization. We provide empirical simulations to justify our theoretical findings. Copyright 2024 by the author(s)
Adversarial imitation learning (AIL) has stood out as a dominant framework across various imitation learning (IL) applications, with Discriminator Actor Critic (DAC) (Kostrikov et al., 2019) demonstrating the effectiv...
Predicting financial markets and stock price movements requires analyzing a company's performance, historic price movements, industry-specific events alongside the influence of human factors such as social media a...
详细信息
We present a novel approach for test-time adaptation via online self-training, consisting of two components. First, we introduce a statistical framework that detects distribution shifts in the classifier's entropy...
With a focus on computationally intensive, distributed, and parallel workloads, scheduling in mixed-criticality distributed systems presents significant challenges due to shared memory and resources, as well as the di...
详细信息
In the field of adversarial games, existing decision-making algorithms primarily rely on reinforcement learning, which can theoretically adapt to diverse scenarios through trial and error. However, these algorithms of...
详细信息
In recent years, the advancement of AI has been primarily driven by neural networks, which, despite their success, pose challenges in terms of explainability and high-power consumption. Genetic Programming (GP) offers...
详细信息
The main control objective for the quad-rotor system is the attitude and position tracking control which is accomplished in this article using the backstepping fractional-order sliding mode control approach combined w...
详细信息
暂无评论