README.md

metadata

license: apache-2.0
tags:
  - robotics
  - reinforcement-learning
  - dexterous-manipulation
  - mujoco

Cyclops Reorient — Reward Weight Ablations

Default Reward Parameters

Index	Name	Default Weight
0	orientation_alignment	30.0
1	hand_pose	0.2
2	action_rate	-0.5
3	torque	0
4	tip_slide	-0.3
5	cage_escape	-100.0
6	finger_collision	-1.0
7	hold_escalation	11.4
8	palm_detach	0.5

Success threshold: 0.2 rad (~11.5°) | Hold steps: 5 | Goal switch delay: 20 steps

Ablation Runs

Run	SLURM Job	orientation_alignment	hand_pose	action_rate	torque	tip_slide	cage_escape	finger_collision	hold_escalation	palm_detach
1 — default	29156044	30.0	0.2	-0.5	0	-0.3	-100.0	-1.0	11.4	0.5
2 — action_rate -1.0	29156071	30.0	0.2	-1.0	0	-0.3	-100.0	-1.0	11.4	0.5
3 — action_rate -0.3	29156287	30.0	0.2	-0.3	0	-0.3	-100.0	-1.0	11.4	0.5
4 — orientation_alignment 60.0	29156420	60.0	0.2	-0.5	0	-0.3	-100.0	-1.0	11.4	0.5
5 — hold_escalation 20.0	29156999	30.0	0.2	-0.5	0	-0.3	-100.0	-1.0	20.0	0.5
6 — tip_slide -0.1	29157532	30.0	0.2	-0.5	0	-0.1	-100.0	-1.0	11.4	0.5
7 — cage_escape -50.0	29157599	30.0	0.2	-0.5	0	-0.3	-50.0	-1.0	11.4	0.5
8 — cage_escape -50.0 orientation_alignment 60.0	29158153	60.0	0.2	-0.5	0	-0.3	-50.0	-1.0	11.4	0.5
9 — cage_escape -20.0 orientation_alignment 100.0	29200037	100.0	0.2	-0.5	0	-0.3	-20.0	-1.0	11.4	0.5
10 — cage_escape 0.0 orientation_alignment 100.0	29205477	100.0	0.2	-0.5	0	-0.3	0.0	-1.0	11.4	0.5
11 — action_rate -0.1	29211473	30.0	0.2	-0.1	0	-0.3	-100.0	-1.0	11.4	0.5
12 — action_rate 0	29211519	30.0	0.2	0	0	-0.3	-100.0	-1.0	11.4	0.5
13 — action_rate 0 cage_escape -20.0 orientation_alignment 100.0	29211541	100.0	0.2	0	0	-0.3	-20.0	-1.0	11.4	0.5
14 — action_rate 0 cage_escape -20.0 orientation_alignment 100.0 tip_slide 0	29211543	100.0	0.2	0	0	-0.3	-20.0	-1.0	11.4	0.5

Bold values indicate parameters that differ from the default.

TIP_COLLISION_GEOMS modified to include "distal_link_(center|left|right)_col" | 15 — default | 29247996 | 30.0 | 0.2 | -0.5 | 0 | -0.3 | -100.0 | -1.0 | 11.4 | 0.5 | | 16 — action_rate 0 cage_escape -20.0 orientation_alignment 100.0 | 29248859 | 100.0 | 0.2 | 0 | 0 | -0.3 | -20.0 | -1.0 | 11.4 | 0.5 |