🎤 Speakers: Claire Tomlin · Amy Zhang · Michael Lutter · Zhongyu Li 👥 Organizers: Jonas Frey · Rosario Scalise · Bike Zhang · Siming He · Aayam Kumar Shrestha · Skand Peri · Yue Hu · Haoran Xu · Arjun Krishna · Dinesh Jayaraman · Shankar Sastry · Susmit Jha
0
0
3
129
We especially welcome early-stage work, new problem formulations, negative results, and lessons from real deployments. 📅 Deadline: Oct. 8, 2026 🏆 Best Paper: $1,000 📍 CoRL 2026, Austin, TX
1
0
2
120
We’re especially interested in three questions: 1️⃣ How should we design simulation across fidelity, diversity, and abstraction for field deployment? 2️⃣ How do we move from lab metrics to field guarantees? 3️⃣ How can robots adapt safely to distribution shifts we didn’t anticipate?
1
0
0
118
Sim-to-real has become a major success story in robot learning. But “real” often means one robot, in one lab, under conditions chosen by researchers. What happens when the same system must work across robots, sites, and unanticipated conditions? That’s sim-to-real-to-field.
1
0
1
184
🎉 Workshop on Sim-to-Real-to-Field: The Science of Transfer is coming to #CoRL2026 in Austin! Our CFP is now open — with a $1,000 Best Paper Award generously sponsored by @tensrinc .
🔗 sim2real2field-corl26.github.io
🔗 sim2real2field-corl26.github.io
3
4
18
8,324
🎤 Speakers:
Claire Tomlin · Amy Zhang · Michael Lutter · Zhongyu Li
👥 Organizers:
Jonas Frey · Rosario Scalise · Bike Zhang · Siming He · Aayam Kumar Shrestha · Skand Peri · Yue Hu · Haoran Xu · Arjun Krishna · Dinesh Jayaraman · Shankar Sastry · Susmit Jha
Claire Tomlin · Amy Zhang · Michael Lutter · Zhongyu Li
👥 Organizers:
Jonas Frey · Rosario Scalise · Bike Zhang · Siming He · Aayam Kumar Shrestha · Skand Peri · Yue Hu · Haoran Xu · Arjun Krishna · Dinesh Jayaraman · Shankar Sastry · Susmit Jha
0
0
0
90
We especially welcome early-stage work, new problem formulations, negative results, and lessons from real deployments.
📅 Deadline: Oct. 8, 2026
🏆 Best Paper: $1,000
📍 CoRL 2026, Austin, TX
📅 Deadline: Oct. 8, 2026
🏆 Best Paper: $1,000
📍 CoRL 2026, Austin, TX
1
0
0
92
We’re especially interested in three questions:
1️⃣ How should we design simulation across fidelity, diversity, and abstraction for field deployment?
2️⃣ How do we move from lab metrics to field guarantees?
3️⃣ How can robots adapt safely to distribution shifts we didn’t anticipate?
1️⃣ How should we design simulation across fidelity, diversity, and abstraction for field deployment?
2️⃣ How do we move from lab metrics to field guarantees?
3️⃣ How can robots adapt safely to distribution shifts we didn’t anticipate?
1
0
0
40
Sim-to-real has become a major success story in robot learning. But “real” often means one robot, in one lab, under conditions chosen by researchers.
What happens when the same system must work across robots, sites, and unanticipated conditions? That’s sim-to-real-to-field.
What happens when the same system must work across robots, sites, and unanticipated conditions? That’s sim-to-real-to-field.
1
0
0
98
🎉 Workshop on Sim-to-Real-to-Field: The Science of Transfer is coming to #CoRL2026 in Austin!
Our CFP is now open — with a $1,000 Best Paper Award generously sponsored by @TenSR.
🔗 sim2real2field-corl26.github.io
Our CFP is now open — with a $1,000 Best Paper Award generously sponsored by @TenSR.
🔗 sim2real2field-corl26.github.io
2
3
11
522
Siming He retweeted
Could long-context architectures “find all contradictions” in a science literature? Not yet! 🧵
We study a new class of "high-complexity” tasks whose difficulty scales quadratically with corpus size (as opposed to linearly), reversing common LCLM decisions! (block-sparse s…)
We study a new class of "high-complexity” tasks whose difficulty scales quadratically with corpus size (as opposed to linearly), reversing common LCLM decisions! (block-sparse s…)
15
70
340
61,141
@TaraSadjadpour @ckwolfeofficial @HaozhiQ @LeaMue27 @JitendraMalikCV It has been great to work on this project!
1
0
8
791
Siming He retweeted
One human demonstration. Any multi-fingered hand. Zero-shot sim-to-real visuomotor policy.
morphometricimitation.github.io
Collaborators: @he_siming @ckwolfeofficial @HaozhiQ @LeaMue27 Shankar Sastry, Claire Tomlin, @JitendraMalikCV
morphometricimitation.github.io
Collaborators: @he_siming @ckwolfeofficial @HaozhiQ @LeaMue27 Shankar Sastry, Claire Tomlin, @JitendraMalikCV
19
32
266
85,843
🧑🌾 6/6 It was amazing to work on this project with Leo Huang, Adam Lilja, Fabio Hübel, Jonas Frey, Marco Pavone (@drmapavone), Shankar Sastry, Jitendra Malik (@JitendraMalikCV), and Claire Tomlin.
0
0
1
42
🌻 4/6 This structured relational memory makes retrieval scalable, interpretable, and robust, without relying on an end-to-end VLM to reason over the robot’s full history.
FARM achieves state-of-the-art performance on referential object retrieval.
FARM achieves state-of-the-art performance on referential object retrieval.
1
0
0
20
🚜 3/6 As the robot moves, FARM constructs an open-vocabulary object memory in real time at 5–10 Hz.
At query time, a VLM converts language into object symbols and spatial relations, which are then grounded explicitly in the memory.
At query time, a VLM converts language into object symbols and spatial relations, which are then grounded explicitly in the memory.
1
0
0
27
🦮 2/6 A robot exploring for hours cannot repeatedly reason over its entire observation history.
FARM instead builds a compact cognitive map: what was seen, where it was seen, and how objects relate, so the robot can retrieve the right object later from natural language.
FARM instead builds a compact cognitive map: what was seen, where it was seen, and how objects relate, so the robot can retrieve the right object later from natural language.
1
0
0
124