skip to Main Content
Reflections From Hacarus Tech Meet-Up In Taiwan

For this week’s focus Ryan Ho, one of our new hires from Taiwan shares his initial impressions from a tech meet-up we held in Taipei in the beginning of the autumn this year. His account of the event follows below:

Hi, my name is Ryan, I joined Hacarus data science team this October, but I’ve been a big fan of the Hacarus’s explainable and lightweight AI since I visit the website. I’ve been looking into the “deep” learning and issues related to the black box problem  for years, so when I have a chance to learn about Hacarus team and their sparse modeling based approach that delivers explainable AI, I see great potential for me to pick up new skills and further my data scientist career.

My first real meeting with the team was just ahead of the tech meet-up that I attended both as new hire and as a member of the audience. 

During the event, Hacarus CTO Takashi talked about sparse modeling and Hacarus’s ambition about next generation explainable AI. One of the backend engineers from the Phillipine team, Ninz talked about the communication between data science  and backend teams. 

Richie Tsai, the Chief Operating Officer of AI Academy, which is the famous AI talent nurturing organization in Taiwan also joined to share some insights on the latest applications for  AI technology. He was very optimistic about the collaboration between the AI industry in Taiwan and Japan, spanning across business partners, technical collaboration and talent nurturing. I see myself as a living example of the collaboration between and hope I could contribute to it.

Afterwards in open discussion, I can feel the excitement and curiosity from our Taiwan audience, who comes from varied industrial background, such as brain science, semiconductor and even natural language processing. But also, I can feel the confusion about the black box model from our audience, since they have great domain knowledge about the industry, however black-box model can’t offer a good explanation about what it have learned and how it’s going to predict.   

Following the event, I felt even more happy to join Hacarus! 

Traditional Chinese Version: 

嗨 ,我是在今年10月加入Hacarus的Ryan,但我早在加入Hacarus之前便是他們explainable 且lightweight AI的粉絲。在我知道這家公司之前,我也累積了多年在深層學習的經驗,並不斷面臨blackbox的問題。因此我從公司網站了解到Hacarus的explainable, lightweight AI目標後,我認為這是我身為資料科學家的下一個目標。

我很高興有機會在正式加入前,參與他們今年九月在台北的招募活動。

我們的CTO 染田貴志說明了Hacarus 如何藉由Sparse modeling達到explainable, lightweight AI的目標。同時我們的後端工程師 Ninz 分享了他對後端工程師與資料科學家之間交流的看法。在活動中我們也邀請到台灣人工智慧學校的Richie Tsai 來為我們分享他對目前全球AI技術發展的看法。Richie Tsai 非常看好台灣未來在AI方面的發展,也對台灣的人才充滿信心,同時也樂見於台灣與日本之間各方面的交流,不管身為商業夥伴,技術上交流的對象,甚至是共同對AI人才的培育。我想我在這方面成為了一個很好的例子,並且也希望能在此盡一份心力。

在會後的討論當中,從在場來自各界的專家中,我可以感覺到他們對explainable lightweight AI充滿了期待,同時我也看到了Black box模型對各個領域的專業人士仍是一個謎團。

在我們的聽眾中,有來自腦科學,半導體,電子零件的研究人員,在各個領域有著豐富的研究經驗 ,但black box模型至今仍無法對各個決策或預測提供一個讓人信服的解釋。

相較於目前深層學習的榮景,在許多領域或應用場景下黑盒狀態仍是一個無法接受的解決方案。因此我更加期待在加入Hacarus後,接觸到各方面醫療或製造業的應用場景後,能夠做出一些改變。