| viXra.org > Artificial Intelligence > 2209 |
 |
 |
| viXra.org > Artificial Intelligence > 2209 |
|
|
[7] viXra:2209.0153 [pdf] submitted on 2022-09-27 06:59:48
Authors: Meng Cao, Ji Jiang, Qichen Ye, Yuexian Zou
Comments: 4 Pages. Technical Report for WAIC Challenge of Financial QA under Market Volatility
This technical report presents the 1st winning model for Financial Community Question-and-Answering (FCQA), which is a task newly introduced in the Challenge of Financial QA under Marker Volatility in WAIC 2022. FCQA aims to respond to the user’s queries in the financial forums with the assistance of heterogeneous knowledge sources. We address this problem by proposing a graph transformer based model for the efficient multi-source information fusion. As a result, we won the first place out of 4278 participating teams and outperformed the second place by 5.07 times on BLUE.
Category: Artificial Intelligence
[6] viXra:2209.0146 [pdf] submitted on 2022-09-28 02:18:16
Authors: Clark M. Thomas
Comments: 6 Pages.
Sentience once mostly referenced human feelings.Now it also points to any "intelligent feelings," with no clear definition emerging. Species inside Earth’s biosphere manifest advanced sentience far beyond everyday awareness. Complex sentience has been critical for complex evolution. Will android robots develop advanced consciousness? Could advanced AI transcend human social sentience, in addition to being super-smart computers? How might UFOs interface with our emerging matrix of advancing technology and imminent ecological disaster?
Category: Artificial Intelligence
[5] viXra:2209.0089 [pdf] submitted on 2022-09-13 02:31:50
Authors: Michael Blackwell, Qing Tian
Comments: 5 Pages.
The goal of this project was to develop a fully convolutional neural network (FCNN) capable of identifying the region of interest (ROI) in dermatoscopic images. To achieve this goal, a U-Net style model was developed for this task and enhanced with an attention module which operated on the extracted features. The addition of this attention module improved our model's semantic segmentation performance and increased pixel-level precision and recall by 4.0% and 4.6%respectively. The code used in thie paper can be found on the project github page: https://github.com/Michael-Blackwell/CapstoneProject
Category: Artificial Intelligence
[4] viXra:2209.0082 [pdf] submitted on 2022-09-14 00:41:01
Authors: G. Torimaru
Comments: 2 Pages.
I explain why consciousness is non-algorithmic, and strong AI cannot come true, and reinforce Penrose's argument.
Category: Artificial Intelligence
[3] viXra:2209.0069 [pdf] replaced on 2022-11-17 03:10:13
Authors: Ait-Taleb Nabil
Comments: 14 Pages.
In this paper, we will propose a method for learning signals related to a data frame $D_{1}$. The learning algorithm will be based on the biggest entropy variations of a Bayesian network. The method will make it possible to obtain an optimal Bayesian network having a high likelihood with respect to signals $D_{1}$. From the learned optimal Bayesian network, we will show what to do to infer new signals $D_{2}$ and we will also introduce the prediction quality $Delta_{CR}$ allowing to evaluate the predictive quality of inferred signals $D_{2}$. We will then infer a large number (10000) of candidate signals $D_{2}$ and we will select the predictive signals $D_{2}^{*}$ having the best prediction quality. Once the optimal signals $D_{2}^{*}$ obtained, we will impose the same order of scatter (computed from the Mahalanobis) to the points of signals $D_{2}^{*}$ as of signals $D_{1}$.
Category: Artificial Intelligence
[2] viXra:2209.0007 [pdf] submitted on 2022-09-02 01:35:30
Authors: Chengkai Guo
Comments: 4 Pages.
In this paper, we first review some of the innovations in modeling mentalizing.Broadly, this involves building models of computing World Model and Theory of Mind(ToM). A simple framework, FaithNet, is then presented with concepts like persistence, continuity, cooperation and preference represented as faith rules.FaithNet defines a generative model that can sample faith rules. Our FaithNet utilize a general-purpose conditioning mechanism based on cross-attention, offering computations that best explain observed real-world events under a Bayesian criterion.
Category: Artificial Intelligence
[1] viXra:2209.0005 [pdf] submitted on 2022-09-01 01:01:30
Authors: Mojtaba Heydari, Frank Cwitkowitz, Zhiyao Duan
Comments: 8 Pages. The 22rd International Society for Music Information Retrieval Conference (ISMIR 2021)
The online estimation of rhythmic information, such as beat positions, downbeat positions, and meter, is critical for many real-time music applications. Musical rhythm comprises complex hierarchical relationships across time, rendering its analysis intrinsically challenging and at times subjective. Furthermore, systems which attempt to estimate rhythmic information in real-time must be causal and must produce estimates quickly and efficiently. In this work, we introduce an online system for joint beat, downbeat, and meter tracking, which utilizes causal convolutional and recurrent layers, followed by a pair of sequential Monte Carlo particle filters applied during inference. The proposed system does not need to be primed with a time signature in order to perform downbeat tracking, and is instead able to estimate meter and adjust the predictions over time. Additionally, we propose an information gate strategy to significantly decrease the computational cost of particle filtering during the inference step, making the system much faster than previous sampling-based methods. Experiments on the GTZAN dataset, which is unseen during training, show that the system outperforms various online beat and downbeat tracking systems and achieves comparable performance to a baseline offline joint method.
Category: Artificial Intelligence
| Third-Party Links: |
|