Summary of DATE 2018 (Design Automation and Test in Europe) - 小结DATE2018会议

---

In this year, DATE conference received 766 paper submissions, which means very bad, normally there are always about 1,000 submissions. What is the reason? I think everyone in academic field knows the reason, but nobody want to reveal it. Out of such total 766 submissions, a large share(39%) has come from authors in Europe, 26% of submissions are from America, 34% from Asia (this number increases so much). And finally, there are 185 papers(24%) were selected for regular presentation and 85 additional ones for interactive presentation.

What is the meaning of DATE, D for Design methods and tools, A for Application design, T for test, reliability and robustness, and E for Embedded system and Cyber-physical system. So the hot topics in this year DATE, for sure the machining learning is the hottest topic in very ACM conference, while beside it, such as emerging technologies, low power challenges and approximate computing for IoT, security-aware design in cyber-physical systems and so on.

---

If the traditional EDA and testing field is dead.... what is the next? Maybe we can see from the filed where those famous EDA and testing professor working right now. They are system-level design of cyber-physical system and security-aware design of cyber-physical system. However, in my opinion, this is not enough. I noticed the BMW group and Volkswagen AG are focusing autonomous driving for the next 100 years, the semiconductor companies are concentrating on FDSOI to overcome the Moore law, also 5G and IoT security are on their way.

---

---

But what is the most promising field right now? I think is quantum computing, every sub-field can serve for quantum computing, while it need too much knowledge of mathematics and physics. I think quantum computing right now just like machining learning in 10 years ago. It's a perfect time to enter.

---

Concerning my own research topics, low power design. After this week in DATE conference, I think I also need change a little bit direction, such as low power design for heterogeneous cyber-physical system or low power design for autonomous systems. The traditional EDA low power design for CMOS is dead without any doubt... Most of the paper just used the methodology from 20 years ago then applied it in another new application. Such kind of reusing is good or not? I don't have answer.

---

For the next 100 years? There are almost more than 200 presentations in these 5 days. The ones talked about self-awareness for autonomous system and electric vehicle. Share with you one video from BMW group used during the presentation did by BMW group in this year DATE.

---

---

今年DATE大会收到766份论文，这个数字其实并不理想，通常总会有大约1,000份论文投稿。 是什么原因？ 我认为学术界的每个人都知道原因，但没有人想说穿。 在这766份投稿中，大部分（39％）来自欧洲，其中26％来自美国，34％来自亚洲（比去年增加了很多）。 最后，有185篇论文（24％）被选中作为长论文发表，另外还有85篇（海报）入选。

如果传统的EDA和测试领域已经死亡......那么下一步是什么？ 除了机器学习，模糊计算，也许我们可以从那些高声望的EDA和测试领域的教授现在工作的方向看到。 短时间内的下一个阶段引人注意的领域system-level design of cyber-physical system and security-aware design of cyber-physical system。但这只是短时间的， 我注意到宝马集团和大众汽车集团在未来100年内都将重点放在自主驾驶核电动汽车上，他们将汽车打造成一个庞大的计算平台，同时各个半导体公司正在专注于FDSOI以克服摩尔定律，通信方面的5G和物联网安全也都是热门方向。

那长远些的方向是什么呢？今年DATE大会上有好几场关于量子计算的special session. 现在的量子计算就像15年前的机器学习一样，绝对的非常好的入场时间，只是这需要很扎实的数学和物理知识。

Autonomous system 是最吸引我本人的地方，特别是 self-awareness for autonomous system。 如果现在的量子计算是15年前的机器学习，花15年的时间，赶上它爆发的时候。。。还是要先好好了解一下。PS， 硬件物理攻击，硬件安全方面今年也特别热。

本文由 @victory622 于steemit上首发。
谢谢您的阅读！
The original post belong to @victory622
Thanks for your reading !
Please feel free to upvote, comment and follow me @victory622