Fast convolution algorithms, including Winograd and FFT, can efficiently accelerate convolution operations in deep models. However, these algorithms depend on high-precision arithmetic to maintain inference accuracy, which conflicts with the model quantization. To resolve this conflict and further improve the efficiency of quantized convolution, we proposes SFC, a new algebra transform for fast convolution by extending the Discrete Fourier Transform (DFT) with symbolic computing, in which only additions are required to perform the transformation at specific transform points, avoiding the calculation of irrational number and reducing the requirement for precision. Additionally, we enhance convolution efficiency by introducing correction terms to convert invalid circular convolution outputs of the Fourier method into effective ones. The numerical error analysis is presented for the first time in this type of work and proves that our algorithms can provide a 3.68x multiplication reduction for 3x3 convolution, while the Winograd algorithm only achieves a 2.25x reduction with similarly low numerical errors. Experiments carried out on benchmarks and FPGA show that our new algorithms can further improve the computation efficiency of quantized models while maintaining accuracy, surpassing both the quantization-alone method and existing works on fast convolution quantization.

Emotions play an essential role in human communication. Developing computer vision models for automatic recognition of emotion expression can aid in a variety of domains, including robotics, digital behavioral healthcare, and media analytics. There are three types of emotional representations which are traditionally modeled in affective computing research: Action Units, Valence Arousal (VA), and Categorical Emotions. As part of an effort to move beyond these representations towards more fine-grained labels, we describe our submission to the newly introduced Emotional Reaction Intensity (ERI) Estimation challenge in the 5th competition for Affective Behavior Analysis in-the-Wild (ABAW). We developed four deep neural networks trained in the visual domain and a multimodal model trained with both visual and audio features to predict emotion reaction intensity. Our best performing model on the Hume-Reaction dataset achieved an average Pearson correlation coefficient of 0.4080 on the test set using a pre-trained ResNet50 model. This work provides a first step towards the development of production-grade models which predict emotion reaction intensities rather than discrete emotion categories.

Emerging applications such as Deep Learning are often data-driven, thus traditional approaches based on auto-tuners are not performance effective across the wide range of inputs used in practice. In the present paper, we start an investigation of predictive models based on machine learning techniques in order to optimize Convolution Neural Networks (CNNs). As a use-case, we focus on the ARM Compute Library which provides three different implementations of the convolution operator at different numeric precision. Starting from a collation of benchmarks, we build and validate models learned by Decision Tree and naive Bayesian classifier. Preliminary experiments on Midgard-based ARM Mali GPU show that our predictive model outperforms all the convolution operators manually selected by the library.

This paper describes a game prototype called "SHRDLU" that explores the concept of designing a game around the ideas behind Winograd's original SHRDLU system. We briefly describe the main gameplay, as well as the natural language and inference architecture used by game NPCs.

Abstract--Recent literature found that convolutional filters into a fractional number field, which is done by neural networks (CNN) with large filters perform well in multiplying the feature maps and filters with some fixed some applications such as image semantic segmentation. These matrices are derived from a Vandermonde matrix, of which the value of Winograd transformation helps to reduce the number of entry numbers grow exponentially with the matrix size. Thus, multiplications in a convolution but suffers from multiplying the data with a large number may make the numerical instability when the convolution filter size gets computation overflow, and dividing the data with a large large. This work proposes a nested Winograd algorithm number makes the computation suffer from quantization error. Compared with the state-of-art OLA-Winograd algorithm, the proposed algorithm Compared with FFT, the Winograd algorithm appears to reduces the multiplications by 1.41 to 3.29 times for be more popular in recent CNN accelerators since it normally computing 5 5 to 9 9 convolutions.

An interesting concept from literary theory states that if a reader wants to make sense of a text, then he will find an interpretation of that text that is consistent with his own world view, or perhaps more precisely, with his view of the world he supposes the text to concern. Oftentimes, to fulfill such a desire requires the reader to fill gaps in his own knowledge, as well as gaps in the logic or rhetoric of the writer by reading between the lines. In this way, all texts are essentially a dialogue initiated by the writer and continued by the reader, with the reader forming, perhaps erroneously, the intentions of the writer. Upon learning of this concept, I fell enamored with writing poetical nonsense with snippets of text found in books and magazines.ᵃ I was excited by the idea of reader attempting to interpret meaning from my curated words and phrases and by doing so finding his own meaning in the resulting lines; perhaps this excitement is a form of sadism -- I don't know -- but during my cut-and-paste creative process, each poem began to take on a personal meaning to me, so perhaps not.

About halfway through a particularly tense game of Go held in Seoul, South Korea, between Lee Sedol, one of the best players of all time, and AlphaGo, an artificial intelligence created by Google, the AI program made a mysterious move that demonstrated an unnerving edge over its human opponent. On move 37, AlphaGo chose to put a black stone in what seemed, at first, like a ridiculous position. It looked certain to give up substantial territory--a rookie mistake in a game that is all about controlling the space on the board. Two television commentators wondered if they had misread the move or if the machine had malfunctioned somehow. In fact, contrary to any conventional wisdom, move 37 would enable AlphaGo to build a formidable foundation in the center of the board. The Google program had effectively won the game using a move that no human would've come up with. AlphaGo's victory is particularly impressive because the ancient game of Go is often looked at as a test of intuitive intelligence. The rules are quite simple. Two players take turns putting black or white stones at the intersection of horizontal and vertical lines on a board, trying to surround their opponent's pieces and remove them from play.

In 1971, Terry Winograd wrote the SHRDLU program while completing his PhD at MIT. SHRDLU features a world of toy blocks where the computer translates human commands into physical actions, such as "move the red pyramid next to the blue cube." To succeed in such tasks, the computer must build up semantic knowledge iteratively, a process Winograd discovered was brittle and limited. The rise of chatbots and voice activated technologies has renewed fervor in natural language processing (NLP) and natural language understanding (NLU) techniques that can produce satisfying human-computer dialogs. Unfortunately, academic breakthroughs have not yet translated to improved user experiences, with Gizmodo writer Darren Orf declaring Messenger chatbots "frustrating and useless" and Facebook admitting a 70% failure rate for their highly anticipated conversational assistant M. Nevertheless, researchers forge ahead with new plans of attack, occasionally revisiting the same tactics and principles Winograd tried in the 70s. OpenAI recently leveraged reinforcement learning to teach to agents to design their own language by "dropping them into a set of simple worlds, giving them the ability to communicate, and then giving them goals that can be best achieved by communicating with other agents."

Via G. Galilei 5, 21027 Ispra (VA), Italy The period since the last conference in this series has been characterized by the explosive expansion of AI out of the confines of institutions of basic research like university departments into the worlds of industry, business, and government (a development I had long expected). But it seems to me that there are plenty-perhaps an overabundance-of other occasions, other conferences, other workshops, and the like, at which the applications of AI would appropriately be considered. In fact, it is ironic-though perhaps it may be understandable-that precisely now, when the outside world has discovered and started showing its appreciation of AI and its potential, there is a widespread malaise among research workers in the field about the health of their subject. This malaise has to do not only with logistic issues such as the drain of very good people from research into applications, or some of the gross inadequacies of structural and funding support by governments. It has to do also with the very heart and methodology of the subject.

The Winograd Schema Challenge (WSC) was proposed by Hector Levesque in 2011 as an alternative to the Turing test. Chief among its features is a simple question format that can span many commonsense knowledge domains. Questions are chosen so that they do not require specialized knoweldge or training and are easy for humans to answer. This article details our plans to run the WSC and evaluate results. Turing (1950) had first introduced the notion of testing a computer system's intelligence by assessing whether it could fool a human judge into thinking that it was conversing with a human rather a computer.