44 Photos and videos
The Rising Star Award has been announced!
Congratulations to Yining Hong @yining_hong , Rising Star Awardee for Spatial Intelligence, and to finalists Zhiyang Dou, Jiafei Duan @DJiafei , Hezhen Hu, Tiange Xiang @xxtiange , and Junyi Zhang @junyi42 .
The awards are supported by 2077AI, with up to USD 30,000 in research gift funding to the awardee's institution and USD 2,000 in API credits for each finalist, helping early-career researchers further develop promising ideas in spatial intelligence.
Full Rising Star list: e2e3d.github.io/rising_star.…
Join the E2E3D Workshop today: 13:00–18:00 · Room 501
E2E3D Workshop: e2e3d.github.io/index.html
1
14
28
14,761
Come meet 2077AI at #CVPR2026 Booth #716.
Stop by to talk about benchmarks, multimodal data, evaluation, and open research collaboration.
Researchers, engineers, students, and friends are all welcome.
@ZihanWang123 @Rubyzx67 @H7803954325458 @Neutrino_l
3
8
327
Wrapped up the DataMFM Workshop at #CVPR2026!
The workshop brought together discussions on the data foundations of multimodal AI, aiming to help define the foundation of next-generation multimodal data ecosystems.
Congratulations to the DataMFM Challenge winners!
Many thanks to our speakers @RanjayKrishna @liuziwei7 @du_yilun @aagrawalAA, workshop organizers Pengyuan Li from @MITIBMLab @ZexueHe @ZihanWang123 @Rubyzx67 @WenhuChen @ManlingLi_ @RogerioFeris, the challenge organizers @aa_blueshark @HanSineng @H7803954325458, and everyone who joined us!
DataMFM Workshop: datamfm.github.io/
#CVPR2026 #DataMFM #2077AI
5
5
377
Great to see this level of engagement at the E2E3D Workshop @ CVPR in Denver!
The room is at full capacity, with many attendees standing and many more gathered outside the room.
Huge thanks to our speakers @lucacarlone1 @jiajunwu_cs @pesarlin @geopavlakos @drmapavone and all the organizers @zhiwen_fan_ @QianqianWang5 @CongWenyan0320 @ManlingLi_ @_vztu @Neutrino_l @ZihanWang123.
This workshop is sponsored by 2077AI, and we remain committed to advancing frontier research in spatial intelligence and embodied AI.
Workshop: e2e3d.github.io/index.html
#CVPR2026 #2077AI
7
14
1,025
Current VLM pipelines have a major blind spot in long-horizon household tasks: noisy auto-labels and weak spatial grounding.
These errors compound fast, leading to object hallucinations, skipped steps, weak physical reasoning, and loss of task intent over time.
Our #CVPR2026 paper, EgoTL, addresses this data bottleneck:
Paper: arxiv.org/abs/2604.09535
1
4
414
EgoTL-Bench evaluates 6 egocentric capabilities across 3 reasoning layers:
memory-conditioned planning
scene-aware action reasoning
next-action prediction
action recognition
direction recognition
distance estimation
It also evaluates world models on 60-second egocentric rollouts.
1
2
104
Results: EgoTL improves both VLM reasoning and world-model rollouts.
With LoRA fine-tuning on ~1.2k curated Q&A pairs, Qwen2.5-VL-7B surpasses the strongest tested baselines across the reported EgoTL-Bench metrics.
Distance estimation is the clearest gain: 39.45% MRA, nearly 2x the best pre-finetuning setup.
Fine-tuned COSMOS-Predict2 also improves long-horizon rollouts, better following think-aloud instructions while preserving object identity and scene layout.
Project: ego-tl.github.io/
Dataset:huggingface.co/datasets/luuu…
2
70
ChartNet, our paper developed with @IBM and @MIT has been accepted at #CVPR2026.
It introduces 1.5M multimodal chart samples across 24 chart types and 6 plotting libraries.
Core question: how do we train models to connect a chart image with the numbers, code, text, and reasoning behind it?🧵
1
1
8
163
Full paper and dataset:
arxiv.org/abs/2603.27064
huggingface.co/datasets/ibm-…
1
44