没有合适的资源?快使用搜索试试~ 我知道了~
人工智能生成的沉浸式 AIGC 服务投标.pdf
1.该资源内容由用户上传,如若侵权请联系客服进行举报
2.虚拟产品一经售出概不退款(资源遇到问题,请及时私信上传者)
2.虚拟产品一经售出概不退款(资源遇到问题,请及时私信上传者)
版权申诉
0 下载量 79 浏览量
2024-03-29
07:40:14
上传
评论
收藏 3.01MB PDF 举报
温馨提示
试读
5页
人工智能生成的沉浸式 AIGC 服务投标.pdf
资源推荐
资源详情
资源评论
AI-Generated Bidding for Immersive AIGC
Services in Mobile Edge-Empowered Metaverse
Zi Qin Liew
∗
, Minrui Xu
†
, Wei Yang Bryan Lim
†
, Dusit Niyato
†
, Dong In Kim
‡
∗
Alibaba-NTU Singapore Joint Research Institute, Nanyang Technological Univerity, Singapore
†
School of Computer Science and Engineering, Nanyang Technological University, Singapore
‡
Department of Electrical and Computer Engineering, Sungkyunkwan University, Suwon 16419, South Korea
Abstract—Recent advancements in Artificial Intelligence Gen-
erated Content (AIGC) provide personalized and immersive
content generation services for applications such as interactive
advertisements, virtual tours, and metaverse. With the use of
mobile edge computing (MEC), buyers can bid for the AIGC
service to enhance their user experience in real-time. However,
designing strategies to optimize the quality of the services won can
be challenging for budget-constrained buyers. The performance
of classical bidding mechanisms is limited by the fixed rules
in the strategies. To this end, we propose AI-generated bidding
(AIGB) to optimize the bidding strategies for AIGC. AIGB
model uses reinforcement learning model to generate bids for
the services by learning from the historical data and environment
states such as remaining budget, budget consumption rate, and
quality of the won services. To obtain quality AIGC service,
we propose a semantic aware reward function for the AIGB
model. The proposed model is tested with a real-world dataset
and experiments show that our model outperforms the classical
bidding mechanism in terms of the number of services won and
the similarity score.
Index Terms—Artificial intelligence generated content, artifi-
cial intelligence generated bid, budget-constraint bidding
I. INTRODUCTION
Artificial Intelligence Generated Content (AIGC) is a tech-
nology that uses artificial intelligence (AI) algorithms to
automatically create and customize content [1] such as videos,
images, and audio. In the context of the metaverse, AIGC can
be used to create immersive and interactive experiences for
users. For example, AI-generated video can be used to create
personalized and interactive advertisements, virtual tours, or
training sessions. By leveraging AI-generated content, the
metaverse can provide highly engaging and realistic experi-
ences for users, enhancing the overall user experience. Edge-
enabled AIGC generation involves the use of mobile edge
computing (MEC) to process data at the edge of the network,
reducing latency and improving overall user experience. This
approach can be applied to AI-generated content creation,
such as videos, images, and audio. By leveraging MEC, AI
algorithms can generate content in real-time or near real-time,
providing highly engaging and personalized experiences for
users [2]. The use of edge computing can also enhance the
security and privacy of the users’ data, as sensitive information
can be processed locally without the need for transmitting it
to external servers. Overall, edge-enabled AIGC generation
can significantly enhance the capabilities of the metaverse and
provide highly engaging and realistic experiences for users.
To obtain AIGC in real-time, we can adopt real-time bidding
(RTB) mechanisms in which service buyers bid for the AIGC
service in real-time. Budget-constrained bidding is a typical
strategy in RTB where the bidders hope to maximize the total
value of returns under a budget. This strategy could help AIGC
bidders to maximize their returns under budget constraints.
The bidding process can be described as follows. During a
time period, one day for instance, there are AIGC service
opportunities arriving sequentially. The bidder gives a bid
according to the AIGC value and competes with other bidders
in real-time. The bidder with the highest bid has the privilege
to obtain the AIGC service and enjoys the value brought by
the AIGC. In the second price auction, the price is determined
by the second highest bid in the auction. The bidding process
terminates whenever the total cost reaches the bidder’s budget
limit or all the AIGC services have gone through the auction.
The goal of budget-constrained bidding is to maximize the
total value of returns under the budget.
Classical bidding mechanisms such as fixed linear bidding
(FLB) linearly scales the bid with a fixed scaling factor.
Budget Smoothed Linear Bidding [3] combines the fixed linear
bidding with the current budget consumption information.
When the budget left ratio is lower than the time left ratio, the
bid is decreased, otherwise, the bid is increased to consume
more budget. Instead of using fixed rules to scale the bids,
AI-generated bidding (AIGB) uses deep learning networks to
learn from the historical bids and their respective returns to
optimize the bidding strategies for future bids. For example,
model-free Reinforcement Learning (RL) [4] is proposed to
resolve the optimization problem of budget-constrained bid-
ding. Deep neural networks are used to learn the appropriate
reward so that the optimal policy can be learned effectively. A
unified solution [5] is proposed to formulate various demands
as constrained bidding problems and then derive a unified
optimal bidding function to achieve the optimum. RL method
is adopted to dynamically adjust parameters to achieve the
optimum. Most of the AIGB models focus on bidding prob-
lems in display advertising. A few works have addressed the
formulation of AIGB mechanisms in AIGC service. Moreover,
the training process of AIGB for AIGC is challenging because
there are limited real-world datasets with AIGC bidding. To
305979-8-3503-3094-6/24/$31.00 ©2024 IEEE ICOIN 2024
资源评论
百态老人
- 粉丝: 1265
- 资源: 2万+
下载权益
C知道特权
VIP文章
课程特权
开通VIP
上传资源 快速赚钱
- 我的内容管理 展开
- 我的资源 快来上传第一个资源
- 我的收益 登录查看自己的收益
- 我的积分 登录查看自己的积分
- 我的C币 登录后查看C币余额
- 我的收藏
- 我的下载
- 下载帮助
安全验证
文档复制为VIP权益,开通VIP直接复制
信息提交成功