Multi-agent Planning using Visual Language Models

ECAI 2024

Examine a statue by the light of a small lamp.(trial_T20190909_193045_208933)

Ground truth Single Agent with table Multi Agent with table Single Agent with image Multi Agent with image
STEP 1: Hang a right at the round table, turn left at the wall and walk to the wooden table by the window.
STEP 2: Pick up the tall statue at the back right corner of the table.
STEP 3: Turn around and walk to the small table next to the couch.
STEP 4: Turn on the small lamp on the table.
Turn around and walk to the SideTable at position (-0.589, 0.000, 1.227).,
Pick up the DeskLamp from position (1.657, 0.654, -1.605).,
Turn around and walk to the SideTable at position (1.581, 0.003, -1.488).,
Place the DeskLamp on the SideTable.,
Turn around and walk to the Statue at position (1.492, 0.661, -1.338).,
Examine the Statue by the light of the DeskLamp.
Walk to the DeskLamp at position (1.657, 0.654, -1.605).,
Pick up the DeskLamp.,
Turn and walk to the statue at position (-1.276, 1.000, 1.327).,
Place the DeskLamp near the statue.,
Turn on the DeskLamp.,
Examine the statue by the light of the DeskLamp.
1. Walk to the lamp on the tripod stand located to the left of the room.
2. Pick up the lamp from the tripod stand.
3. Walk to the statue on the wooden cabinet located on the right side of the room.
4. Position the lamp to illuminate the statue.
1. Identify the location of the statue: The statue is on the cabinet.
2. Walk to the statue at the cabinet.
3. Identify the location of the small lamp: The lamp is on the table.
4. Walk to the table to reach the small lamp.
5. Pick up the small lamp from the table.
6. Walk back to the cabinet with the statue.
7. Place the small lamp near the statue on the cabinet.
8. Turn on the small lamp to illuminate the statue.
9. Examine the statue by the light of the small lamp.
Architecture