Confessions of a researchaholic

September 7, 2021

Good results

Filed under: Real — liyiwei @ 4:56 pm

I remember (20+ years ago while I was still in school) during a faculty seminar, Mark Horowitz couldn’t attend, and Bill Dally presented his research by skipping all results slides by simply saying “good results!” and called it a day.

As someone coming from the SIGGRAPH-side of research, I always wish that all CHI user study sections can contain just one sentence: happy users and good results! (instead of pages of descriptions and statistics which, honestly, I never read).

July 24, 2021

AR face/body filters

Filed under: Imaginary,Real — liyiwei @ 6:02 pm
Tags: , ,

When one day AR glasses become as ubiquitous as smart phones under shared platforms like Android/iOS, AR face/cloth filters could supplement or even replace our physical makeup and attire, as already happening with virtual cameras during virtual meetings that I have been experimenting with recently.

July 20, 2021

ACM SIGGRAPH new technical publication program survey

Filed under: Real — liyiwei @ 2:02 pm

You can find the survey by Adam Bargteil here.

I have a simpler proposal: revert to the original conference format for SIGGRAPH technical papers (i.e., before year 2002) and give authors the option to submit an extended version of their SIGGRAPH conference papers to ToG as journal papers (with the usual rule of at least 30\% more new materials, for example, most likely for additional analysis and evaluation).
We can then get rid of the current special ToG issues of SIGGRAPH and SIGGRAPH Asia.

July 15, 2021

Fill the board with calculus

Filed under: Real — liyiwei @ 9:46 am
Tags: ,

I heard from this podcast about how Dr. Cecilia Conrad was advised to, instead of making her lectures too clear, “fill the board with calculus” to establish her credential for students who thought she was an “affirmative action hire”.

I know plenty people, my past self included, who are natural board-calculus fillers by overwhelming technical communications such as research papers and presentations with math/algorithm details, and need to reverse what Dr. Cecilia Conrad did.

July 1, 2021

Record for the day

Filed under: Real — liyiwei @ 12:30 pm

June 27, 2021

Mindful drawing

Filed under: Real — liyiwei @ 9:08 am
Tags: ,

After 30+ timed-model-drawing sessions my sketches still have only forms but not actions/gestures, which I realized is caused by mimicking what I see but without really understanding.
Specifically, the first lesson of my first drawing training is to draw exactly what I see so that the mind won’t get in the way, but now I need to get my mind back into the process.

(I was so frustrated that I did a few visceral physical drawings for mental balance.)

The quote “you draw with your mind not your hand” from the Vilppu drawing manual (which is a great book not just for drawing) reminded me of what Stephen King said in the gunslinger:

I do not aim with my hand; he who aims with his hand has forgotten the face of his father.
I aim with my eye.

I do not shoot with my hand; he who shoots with his hand has forgotten the face of his father.
I shoot with my mind.

I do not kill with my gun; he who kills with his gun has forgotten the face of his father.
I kill with my heart.

April 16, 2021

Associative memory

Filed under: Real — liyiwei @ 4:12 pm
Tags: ,

Looking at this dithering technique today brought back memories of my previous work on blue noise sampling and the associative life, almost like traveling back in time.

March 7, 2021

OnAir logo

Filed under: Real — liyiwei @ 11:01 am
Tags: ,

I dreamed about this logo of an ongoing project among existing Adobe apps.

February 15, 2021

What I like to see in everything machine-learning paper

Filed under: Real — liyiwei @ 1:58 pm

Does this problem require ML, or a hand-crafted method would be better (to control, debug, understand, implement, etc.)?

Is this better for a paper or just open source?

How did you come up with the architecture, and why?

Next Page »

Theme: Rubric. Get a free blog at