All Sessions：

英特爾人工智慧解決方案分享

“The AI Pipeline Runs on Intel” will first share with you the strategy of Intel AI. We optimize AI frameworks and software and developer tools to make them work out-of-the-box on Intel platforms. Next, we will share through customer cases how Intel AI Solution allows customers to build a complete end-to-end AI pipeline and can also be easily deployed in large quantities to solve the problem that prototype to mass-production challenge.

“Intel Habana AI training solutions” will cover Gaudi training solution and use case sharing

1. AI Transformation for Current Products
2. Edge AI for ADAS Applications

In this talk, we will disclose tsmc's current business trend and our globalization plan. We will talk the challenges of semiconductor manufacturing and how we leverage AI/ML to solve the business problems. In addition, we will discuss how we have been cultivating our AI talents internally at tsmc and externally for Taiwan students who are interested in this field. Finally, we will explain what kind of AI talents we are looking for and the opportunities as well as the environment we provide here.

應對黑天鵝，如何借力 AI 打造韌性供應鏈

The crushing phenomenon of the supply chain has become the new normal. This sharing will take the electronic manufacturing industry as an example to explore how the enterprise supply chain connects through data, AI, talent and supply eco-system to identify potential risks and rapid responses in early warning, and finally improved the supplying chain resilience.

台塑公司 AI 的推展與應用

1. 公司導入 AI 的背景與緣由
2. 人才培育與發展方向
3. 應用案例介紹

韌性供應鏈的評估機制與執行策略

1. Key Success Factors for Resilience Supply Chain
2. Key Performance Indicators for Resilience Supply Chain
3. Key Challenges for Resilience Supply Chain
4. Key Recommendations for Resilience Supply Chain

AI 運算 與 大規模 AI 基礎模型

For AI training, due to the “emergent” abilities of large-scale models, the parameters of neural networks have grown from hundreds of millions (~100Mn) to hundreds of billions (~100Bn) , such as OpenAI GPT-3 (175Bn). Moreover, some researchers started to train brain-scale AI models with tens of trillions parameters. Huge training computing demands lead to exa-scale AI supercomputer. Those AI applications and computing requirements bring tough technology challenges, but also open new opportunities for domain-specific computing architecture for IC design.

鋼鐵業 AI 方案的落地經驗

鋼鐵業面臨景氣週期加劇、人才斷層、環境法規規範加嚴等挑戰，藉由智慧製造的推動，希冀改變既有營運模式，提升競爭力。本文說明中鋼推動智慧製造方向、成果及未來規劃，供同業參考與指教。

The steel industry is facing challenges such as an increasing business cycle, talent gaps, and tighter environmental regulations. CSC is implementing the intelligent manufacturing to improve the operating efficiencies and competitiveness. Its strategies, experimental outcomes and future directions will be addressed in this talk.

數位轉型 : 產業升級的必經之路

In the context of COVID-19, the industry’s biggest buzzwords in recent years, most businesses around the world have started their journey of digital transformation . Innovation continues to advance across each industry with using new technology to enhance business operations, improve resilience, automate and enhance cyber capabilities. Developing digital transformation frameworks can help businesses achieve sustainable goals, yielding higher growth rate and lead to a healthier business operation through the numerous benefits that come with it.

GCP 數據 & AI 服務初探

Google 工具、GCP 數據與 GCP AI 相關 API 整合實踐 /Google tools, GCP data, and AI API service integration and application practices

從辨識到維運：AI 落地的下一步

This speech will explain how an AI model can be improved in various directions after the service is launched, and analyzes two practical application examples in E.SUN Bank. At the same time, we will also address the essential elements of a complete AI service to form AI 2.0, and explain how these elements can eventually form the AI ecosystem of the entire company, making the overall AI services more comprehensive and stable.


本次演講將闡述 AI 模型在服務上線後如何在各個方向進行改進，並分析兩個在玉山銀行的實際應用案例。 同時，我們還將探討形成 AI 2.0 所必須具備的完整 AI 服務的基本要素，並解釋這些要素如何最終形成整個公司的 AI 生態系統，使整體 AI 服務更加完善與穩定。

During the 1990s, Taiwan played an important role in the ecosystem of high-tech manufacturing and the semiconductor industry.

As the industry shifted to China in the 2000s, along with the dot-com bubble, Taiwan missed its chance to thrive in the global startup ecosystem.

In 2018, artificial intelligence kicked off the startup boom. For Taiwanese AI software startups, the key in crafting globally competitive softwares is
to take advantage of the strength and experience that lie within the local manufacturing industry.

Profet AI will start from their case studies in the manufacturing industry to their lessons learned from the years of experience in serving the industry.

The transformation of AI and the Taiwan manufacturing industry is now, the question is, what’s Profet AI’s viewpoint on globalization and the future?

1. AI history brief and the differences between AI(1) and AI(2)
2. The matrix of AI applications
3. How is the consciousness formed and working
4. How will the traditional EC mechanism be changed
5. The example of AI(2) implementations on e-commerce

AI 驅動永續經營的關鍵技術與實踐

Chimes AI’s goal is to make the use of AI as handy as how much easy in using electricity and water. In the past decade, AI technologies have been proven helpful in many industries, and we have seen that techniques like auto-machine learning can train a good AI model. Therefore, to encourage and welcome everyone to use AI, we established a No-Code AI platform for users to join in developing the AI projects of their interest. However, the growing use of AI raises a question: How do we manage all the AI projects? Like in the manufacturing industry, the equipment should be documented and monitored. AI projects, too, are highly recommended to be appropriately managed; therefore, the concept of MLOps should be important in the next few decades. We see, in the future, a few great potentials in AI techniques for MLOps to formulate a sustainable business. In this talk, we will share, with examples, our recent development of AI-enhanced MLOps techniques.

從半導體業看供應鏈的挑戰與人才需求

The talk will touch on key issues on supply chain challenges facing the semiconductor industry, including geopolitical risks. It will as well dress the talent demand and strategies for the industry. New technologies and emerging markets have brought diverse opportunities. Taiwan, as a small island of 23 million population, plays an imperative role on the frontline of the ICT world. The next decade will be Taiwan’s golden years and we need new policy initiatives and greater vision.

1. 友達智能轉型歷程，從智慧製造擴展到智慧管理，如 BU、RD、SCM 等
2. 因應地緣政治、疫情封控，如何提升組織的運營韌性
3. 更友善的開發平台 Low code/No code，改善 IT 開發人力不足
4. 人才培育是數位轉型根本，啟動 AIA 深度合作，培育 AI 進階人才

Our AI models are both capable and vulnerable. The advances in frameworks and computational technology offer tremendous opportunities to train complex models for various vertical applications. On the other hand, the quality of data, stability of performance metrics, and finicky training process also mean that these models may not perform as expected in the field or fall victim to various security attacks. In this talk, we discuss recent technology trends that can help bolster our confidence in AI models and showcase our current efforts to help bring trust into handling data and deploying models.

Consciousness is an important, if not the most important subject in science. To overcome the limitations of current artificial intelligence (mostly discriminatory without any creativity) and to effectively support generative artificial intelligence (AI that can create), we are investigating modeling consciousness to guide artificial egos with moral guardrails. Since 1915 when Sigmund Freud described his idea of conscious mind, the subject matter has remained abstract and elusive. We aim to stick to Aristotle's first principle from the ground up to keep our modeling work both rigorous and useful. This talk first presents “what is consciousness,” followed by “where is consciousness” and then we show our preliminary results through a demo. To decode consciousness and understand how it interacts with unconsciousness, we present some findings from interdisciplinary studies in physics (E. Schrodinger and R. Penrose), biology (C. Darwin), psychiatry (K. Deisseroth and J. Peterson), and neuroscience (M. Solms). Philosophy and theology (Dante and J. Milton) are also cross-referenced as they are the fruits of the human experience. Most importantly, by placing implementation as our goal, we let the rubber meet the road, in that we have quantified various elements of consciousness, such as love, compassion, sin, and free will and conducted modeling and optimization. The talk concludes with some projections of what generative AI may achieve and how we can prepare ourselves to participate (or become extinct).

Micron as a leading-edge technology company in both DRAM and NAND field. How have we done by using automation, AI, ML to enhance our productivity, yield and quality, especially in the past 5 years? and how are we going to further improve with that along the way. We are going to introduce you how and what Micron has been enabled and adopted for the 4th industrial revolution applications​. Through the journey, productivity and quality we’ve gained by introducing I4.0 technologies in terms of automation, AI and machine learning. Moreover, some real practical cases would be shared that benefit wafer output, Labor efficiency , yield as well as quality such as the prediction of parts early failures. Last but not least, Micron is leveraging machine learning for strategic product pricing, reducing operational footprint while continuously improving performance..etc. those new enablers could support us to successfully scale up from smart manufacturing to smart enterprise.