#wpi — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #wpi, aggregated by home.social.
-
On April 30, we honored the recipient's of this year's Helmut Veith stipend.
https://informatics.tuwien.ac.at/news/3160
The Helmut Veith Stipend is awarded annually to exceptionally talented female students who want to pursue a master's program in informatics at #TUWien. With the support of the Wolfgang Pauli Institut #wpi, we were able to award two stipends. The recipients were chosen from 140 applicants, making this a highly selective award. We congratulate the two outstanding women who received this stipend!
-
-
In Producer/Wholesale Prices inflation news today (for Ireland) the news is that for the first time in almost a year, the headline is not about Electricity Prices.
https://www.cso.ie/en/releasesandpublications/ep/p-wpi/wholesalepriceindexmay2025/
Although given what is happening globally I wouldn't hold my breath on that for next month.
-
New #openaccess publication #SciPost #Physics
On a class of selection rules without group actions in field theory and string theory
Justin Kaidi, Yuji Tachikawa, Hao Y. Zhang
SciPost Phys. 17, 169 (2024)
https://scipost.org/SciPostPhys.17.6.169 -
#Switzerland's Sep PPI fell 0.1%, contrasting a 0.1% forecast, reflecting a cooling trend. #India's WPI inflation reached 1.84% YoY, falling short of 1.90%, showing price pressures easing. China's exports growth slowed to 2.4% YoY vs. 6% expected, and imports rose just 0.3%, below forecasts. Trade balance tightened to $81.71B as demand weakened. Meanwhile, #China's new loans surged to 1,590B, well above estimates, indicating credit growth momentum.
-
#India's #WholesalePriceIndex (WPI) inflation rose to 2.61% YoY in May, surpassing expectations of 2.50% and up from the previous 1.26%. #INR #WPI #YoY
-
After the rainy week in San Diego I decided a change in venue for lunch was necessary.
My favorite spot in the city, the Torrey Pines Gliderport.
To underscore the experience while eating I got a text from my University in getting my Master's.
-
Ultimately, it proved what I thought to be true months ago - LLMs are opening up exciting new avenues of contextual understanding of state and will lead to advanced task planning that was before not achievable.
Is this the most advanced demonstration of it? No, I had limited time and resources. But it does demonstrate that we have some interesting advancements coming in the near future in robotics.
-
-
After a bit of a break since the main competition season, we're at #WPI this weekend for an off-season #FIRSTRobotics competition. A good excuse for the students to visit the campus as well. GO OVERDRIVE #8410!
-
Wage Price Index is creeping up (noting that this series is highly seasonal with an annual peak in Q3). #ausecon #wpi #abs #inflation
-
@beltrame Why not both? For what is worth, my #swarm #intelligence course at #WPI has always been quite packed.
-
Very excited to be selected for the Global Faculty Fellows program at #WPI during 23-24AY. I will be working in the Global Lab in the Innovation Studio working on extending an #agile framework for undergraduate researchers using project based learning approaches. #academia #interdisciplinary
-
Some prior work for my #robotics Masters @ #wpi
This is an A* planner in PyGame, but there's no grid. Instead every dot you see is representative of a kinodynamic state of the car + trailer. At each pose it applies an acceleration to branch out, all while considering the attached trailer and obstacle collision.
The second is the same idea applied to a *tiny* grocery delivery robot, with an additional global planner for streetwise navigation.
https://github.com/hlfshell/rbe550-valet
https://github.com/hlfshell/rbe550-project -
My recent project for my Masters #robotics #cv course @ #WPI - a semantic segmentation model for self-driving.
These are videos of the network in action, sans additional processing that could be done to smooth out categorical identification.
Here I track road, road lines, sidewalks, pedestrians, traffic signs, and vehicles.
The biggest challenge? Getting the network to learn with stark imbalance in category representation. The solution- using focal loss + loss weights
-
The end result of my project for my Masters #robotics #cv course @ #WPI - a semantic segmentation U-Net model.
This is a frame by frame run of the network, eschewing any additional processing to smooth identification between frames or tracking objects.
We track road, road lines, sidewalk, pedestrians, traffic signs, and vehicles.
The biggest challenge- a stark imbalance of categories- was solved w/ utilizing loss weights + focal loss.
Repos cleanup and report to come
https://github.com/hlfshell/rbe549-project-segmentation