Learning from demonstration

Overview

Learning from demonstration (LfD) means a learner picks up a task from examples recorded while someone, usually an expert, performs it. The survey describes it as a way to avoid explicitly planning the motion, because the robot's trajectory is learned from the demonstrations 2105.01767 §5.1. Demonstrations also appear in the sensing and learning part of the survey, where human examples are said to make end-to-end learning more efficient 2105.01767 §3.2.

The other cards use demonstrations in other ways. Ibarz et al. use them to pretrain a game agent, and fit a reward to pairwise preferences 1811.06521 §2.2. GF-VLA turns one human RGB-D video into a scene graph that a language model plans from 2508.05342 §4.1. PerAct trains one language-conditioned policy by behaviour cloning on keyframe actions 2209.05451 §3.2 2209.05451 §3.4. BC-Z trains one image-to-action policy on a real robot across 100 tasks, with demonstrations and human interventions during rollouts, and is told the task by a language command or a human video 2202.02005 §4 2202.02005 §5.1. AnyTeleop focuses on how the demonstrations are recorded, using a camera-based teleoperation system, and then imitates them 2307.04577 §7.1. A frozen large language model can be given demonstrations in its prompt and asked to continue them, with no training 2307.04721 §5. "Showing versus Doing" treats a demonstration as a message from a teacher, and reports that reward recovery works better from trajectories chosen to show the task 2610.14004 §4.3. VINN keeps the demonstration frames and answers each query by averaging the actions of the closest frames, so no policy is trained after the encoder 2112.01511 §3.2. DriveMLM is trained on recorded expert driving with rule-derived labels 2312.09245 §3.4. The paper does not describe its method as imitation, so it is a looser fit than the others. Three more cards use demonstrations to build a reward or a task structure: TemporalOT compares agent frames with expert frames under an order-aware optimal-transport cost 2410.21795 eq. 5, the reward-machine paper infers a task graph from unlabelled videos 2412.10096 §4.3, and Visual IRL learns a reward from human keypoints 2412.11360 §IV.A.

The cards share the idea of a policy or trajectory taken from recorded behaviour, but none gives a formal definition of LfD, and the survey names no standard algorithm 2105.01767 §5.1. A 2023 survey of imitation learning groups the field into behavioural cloning, inverse reinforcement learning, adversarial imitation and imitation from observation 2309.02473 §I. The imitation-learning page covers the supervised form of this idea in more settings.

Variants

Encoding demonstrated trajectories and forces for planning

The robot learns from human demonstrations, and the learned encoding covers both the robot trajectory and the interaction force. The survey attributes this approach to Lee et al. (2015) and presents it as one of the planning methods for deformable objects 2105.01767 §5.1. Lee et al. is cited in the survey but has no card here, so nothing about its method is taken from this page. The survey reports no success rate or error figure for this approach 2105.01767 §5.1. Card: Challenges and Outlook in Robotic Manipulation of Deformable Objects.

Behaviour cloning from keyframe demonstrations

PerAct learns one multi-task policy for 18 simulated RLBench tasks 2209.05451 tab. 1. Its demonstrations are reduced to keyframes: an action is kept when joint velocities are near zero and the gripper state does not change 2209.05451 §3.2. The paper does not state the velocity threshold. The policy predicts a voxel for the gripper translation, a rotation bin per axis, the gripper state and a collision flag, and each head is trained with cross-entropy 2209.05451 §3.4.

With 10 demonstrations per task, the average gain over the 3D U-Net baseline is 1.33x; with 100 it is 2.83x 2209.05451 §4.2 2209.05451 tab. 1. Without the language goal the agent performs at chance 2209.05451 §4.2. Random or fixed-interval keyframes give zero success, so the keyframe filter is part of the method 2209.05451 §4.2. The real-robot results use 5 to 10 test episodes per task, so they are only indicative 2209.05451 tab. 2. Card: Perceiver-Actor: A Multi-Task Transformer for Robotic Manipulation.

Many-task behaviour cloning with a task embedding

BC-Z conditions one image-to-action policy on a 512-dimensional sentence embedding of the task, injected through FiLM layers in a ResNet-18 torso 2202.02005 §5.1 2202.02005 §5.3. The training set has 25,877 demonstrations: 11,108 expert-only and 14,769 collected with HG-DAgger, where a human takes over when the learned policy goes wrong and the corrections are added to the data 2202.02005 §4. The policy was redeployed over 16 iterations 2202.02005 §4. The survey files this kind of human takeover under interactive imitation, which targets the drift of a learner into states the expert never showed 2309.02473 §II.

On held-out tasks, the average success rate is 32% with a language command and 4% with a video command 2202.02005 tab. 3. The authors read the close training results for one-hot (42%) and language (40%) conditioning as a sign that the language space is adequate, so the held-out gap sits in the policy 2202.02005 §6.2 2202.02005 tab. 3. On one task with 1,000 demonstrations, a single-task policy scores 5% and the multi-task language policy scores 52% 2202.02005 tab. 4. Training the same task with 50% manual and 50% HG-DAgger episodes reaches 53%, against 27% for 100% manual demonstrations at the same episode count 2202.02005 tab. 4.

The paper states 29 held-out tasks in the text and its table has 28 rows 2202.02005 §4 2202.02005 tab. 2. The released code covers only the language-conditioned model. Its learning rate and batch size differ from the paper's, and it has no video encoder run_train_bc_gtcond_trajectory.gin model.py. Card: BC-Z.

Nearest-neighbour lookup over demonstration frames

VINN has no policy head. Its encoder is a ResNet-50 fine-tuned with BYOL on the demonstration frames, without action labels 2112.01511 §3.1 2112.01511 §4.3. At test time the query image is embedded, its k nearest demonstration frames are found under Euclidean distance, and their actions are averaged with softmin weights 2112.01511 §3.2. The paper uses k around 10 and reports no gain beyond k = 20 2112.01511 §4.7. The weights are exp(-distance), with no temperature stated 2112.01511 §3.2.

On a real Hello Robot Stretch opening three cabinets, 30 trials in total, VINN opens the door 80% of the time. Behaviour cloning on frozen representations reaches 53.3%, and end-to-end behaviour cloning reaches 0% 2112.01511 tab. I. When the cabinet is partly occluded, VINN keeps 70% with the sign and handle hidden, where the frozen-representation baseline falls to 10%. This comparison uses 10 trials on one cabinet 2112.01511 tab. II. In the unmodified scene the baseline is better, at 90% against 80% 2112.01511 tab. II.

With few demonstrations, the paper says nearest-neighbour methods are competitive or better, and that VINN is clearly better on Stacking and Door Opening below 20 demonstrations. The comparison is shown only as a plot 2112.01511 §4.4 2112.01511 fig. 4. Each query is compared with every demonstration embedding, so cost grows with the number of frames. The paper reports about 0.038 s per query on hardware it does not state, and it names FAISS as a faster search it did not test 2112.01511 §4.8. Card: The Surprising Effectiveness of Representation Learning for Visual Imitation.

Pose matching to demonstrations with neural descriptor fields

This variant trains no policy. A network trained to reconstruction, without keypoint labels, gives each 3D point a descriptor, and the descriptor field is made SE(3)-equivariant with Vector Neurons 2112.05124 §II-A 2112.05124 eq. 5. A rigid body is represented by a set of query points, and its pose descriptor is the concatenation of the descriptors of those points 2112.05124 §II-B 2112.05124 eq. 10. For a new object, a pose is found by minimising the L1 distance between the candidate's pose descriptor and the average demonstration descriptor, from random starts, with Adam 2112.05124 eq. 11 2112.05124 §II-C.

In simulation with 10 demonstrations per task, overall success is 0.87 to 0.91 on upright mugs, bowls and bottles, and 0.58 to 0.78 when the objects are in arbitrary poses. The Dense Object Nets baseline reaches 0.00 to 0.45 2112.05124 tab. I. Taking the activations of all network layers, rather than the last or first layer alone, is the largest factor in the ablation: 0.88 against 0.65 for the last layer on upright mugs 2112.05124 tab. II. On upright mugs, overall success is 0.46 with one demonstration and 0.70 with five 2112.05124 tab. IV.

The method has limits the card records. The real-robot section gives only qualitative examples 2112.05124 §IV-C. The pose search is non-convex, and the paper does not state the number of starts or steps 2112.05124 §II-B. It transfers poses and points, not whole trajectories, and it assumes rigid objects and a static scene 2112.05124 §III 2112.05124 §VI. Card: Neural Descriptor Fields: SE(3)-Equivariant Object Representations for Manipulation.

Collecting demonstrations by camera teleoperation

This variant is a source of demonstrations, not a learner. AnyTeleop uses a camera to watch the operator's hand, detects the hand pose, retargets it to the robot hand joints by optimization, and moves the arm with a GPU motion generator 2307.04577 §4. Its stated aim is to make demonstration collection cheap across many robots and settings 2307.04577 §1. RGB-only input works, with a wrist-position error that the paper does not quantify 2307.04577 §4.1.

In simulation, the imitation learner (DAPG) is trained on 50 demonstrations per task, with three seeds and 100 evaluation trials 2307.04577 §7.1 2307.04577 tab. 3. AnyTeleop's policy scores higher than the baseline built from the earlier system's demonstrations in 5 of 6 settings, and higher than pure RL in all 6 2307.04577 tab. 3. Several gaps are smaller than the standard deviations over three seeds 2307.04577 tab. 3.

On a real XArm6 with an Allegro hand, one camera and ten runs per task, the paper reports a higher success rate than the Telekinesis system on 8 tasks and an equal one on 2 2307.04577 §6.2 2307.04577 tab. 4. The Telekinesis numbers are copied from its own paper, not re-run 2307.04577 tab. 4. Mapping human finger keypoints to robot joints is set by hand for each robot 2307.04577 §4.3. Card: AnyTeleop.

DexPilot is another camera teleoperation system in the cards. Four RGB-D cameras track the operator's bare hand, and a fingertip-vector cost maps it onto a robot hand with different joints 1910.03135 §7.1. The card describes no policy learned from the recorded operator motion. Its learned networks are perception models, trained on point clouds labelled by a tracker that was seeded with glove data 1910.03135 §6. The reported results are task completion times on 15 tasks, with 9 failed trials out of 150 1910.03135 app. completion times. This is therefore a source of control, not evidence about imitation. Card: DexPilot.

Low-cost glove hardware for recording demonstrations

GEX is also a source of demonstrations, not a learner. It is a teleoperation kit: a three-finger hand with 11 motor-driven joints, and an exoskeleton glove with 12. The glove's fingertip keypoints are retargeted to the hand's joints by optimization 2506.04982 §3.2. The paper says teleoperation for collecting demonstrations is hard with many-DoF hands, and that its aim is a cheap kit for more labs 2506.04982 §1. The author-stated cost is USD 600 per device, against above USD 16,000 for the Allegro hand 2506.04982 §1. The paper reports no success rates, latency or baselines. Its results are one real paper-cup grasp and three simulated tasks, shown qualitatively 2506.04982 §4 2506.04982 §4.1. The paper gives no learning result, so this variant has no measured effect on a policy. The retargeting method in the released code is DexPilot from an external library, not the optimization written in the paper gx10_right_dexpilot.yml. The force-feedback part described in the paper is not in the released code 2506.04982 §3.3. Card: GEX.

Expert driving logs for a language-model planner

DriveMLM fine-tunes a multimodal language model to output one speed decision and one path decision per step, plus an explanation. A motion planner turns the decision into controls 2312.09245 §3.1. Its training data are 280 hours of CARLA driving on 50k routes 2312.09245 §4.1. Experts, human drivers or agents, drive the challenging scenarios. The decision labels come from hand-written rules applied to the expert trajectories, and the explanations are written from scene elements and refined by annotators 2312.09245 §3.4.

In closed loop on CARLA Town05 Long, the paper reports a driving score of 76.1 against 71.4 for Apollo, and 0.96 miles per intervention against 0.76 2312.09245 tab. 4. These are author-reported results from one route set. The paper does not say whether the baselines were re-run or copied 2312.09245 §4.4. The data are available only on request, and no code is released 2312.09245 Declarations. The paper does not state how the decision definitions are written, so the rule set cannot be rebuilt from it 2312.09245 §3.2 2312.09245 fig. 4. Card: DriveMLM.

Scene graphs from one human video, planned by a language model

GF-VLA is a dual-arm system that learns a block-assembly task from one human RGB-D demonstration. It finds hand and object contacts with entropy and mutual information over positions, and writes them as a scene graph with keyframes 2508.05342 §4.1 2508.05342 §3.1. The language-model head writes a plan of subtasks, each with a reason and a check, and the action head emits Cartesian poses and gripper states at 5 Hz for the two arms 2508.05342 §4.2 2508.05342 §4.3. The two heads share one 7B LLaMA 2 backbone with a separate LoRA adapter each 2508.05342 §4.4.

Reported results are for the paper's own tasks. Block-assembly grasp success is 0.94 and placement success 0.89 on average over 120 trials 2508.05342 tab. 1. Transfer from one demonstration reaches 0.90 task success over six variants 2508.05342 tab. 3, against 0.79 and 0.76 for GraspVLA on letter and tower tasks 2508.05342 tab. 4. Removing the scene graph cuts letter and tower success to 0.48 and 0.35 2508.05342 tab. 2. Each transfer variant has 10 trials, but the table reports values such as 0.87 and 0.93, which cannot be counts out of 10, and the baseline runs are not described 2508.05342 §5.4 2508.05342 tab. 3. No code is released. Card: Information-Theoretic Graph Fusion with Vision-Language-Action Model for Policy Reasoning and Dual Robotic Control.

Demonstrations to bootstrap end-to-end reinforcement learning

End-to-end reinforcement learning maps sensor data straight to robot actions. The survey says human demonstrations can make this more efficient, and gives one example: Matas et al. (2018) trained an improved deep deterministic policy gradient method with 20 demonstrations to manipulate cloth [arxiv:2105.01767 §3.2]. The survey notes that such settings often need a hand-designed cost or reward, which is the reason for the next variant 2105.01767 §3.2. The number of 20 demonstrations and the task come from the survey's description of that work, not from the original paper.

Demonstrations with a reward learned from preferences

Here the demonstrations are one of two feedback channels. A game agent is first pretrained on a few expert episodes, and a reward model is then fitted to pairwise human or synthetic preferences over 25-step clips, and the agent learns from that reward 1811.06521 §2.1 1811.06521 §2.4. The demonstrations stay in the replay buffer for the whole run 1811.06521 §2.3. Most labels are synthetic, generated from the game score with a 10% annotator-noise model; human labels are reported separately, with a 5 to 10% mislabelling rate 1811.06521 §3.1 1811.06521 app. A.3. The authors report that the agent beats imitation in 7 of 9 Atari games, with about 1000 comparisons 1811.06521 §4 1811.06521 abstract. The feedback has to keep arriving: with a frozen reward model, scores fell in three games as the agent found loopholes 1811.06521 §3.2 1811.06521 fig. 4. Card: Reward learning from human preferences and demonstrations in Atari.

Recovering the reward from demonstrations

When the reward is hand-designed and hard to get right, the survey suggests using the demonstrations to recover the reward with inverse reinforcement learning 2105.01767 §3.2. The survey gives no example result for this variant. The 2023 survey sorts these methods by how they handle the fact that many rewards fit the same behaviour: maximum margin, maximum entropy and Bayesian approaches 2309.02473 §III. It notes that IRL needs repeated interaction with the environment to estimate the reward, which can be costly or unsafe in high-risk settings 2309.02473 §III. It reports no experiments of its own.

"Showing versus Doing" models the case where the demonstrator is teaching. A teacher chooses trajectories that make the true reward easiest for an observer to infer 2610.14004 §2.2 2610.14004 eq. 4. The paper applies this choice to whole trajectories, and sets the observer to assume a demonstrator who is doing the task with a softmax policy 2610.14004 §2.3 2610.14004 eq. 5. In two small grid tasks on Mechanical Turk, with 59 and 58 analysed participants, people asked to show the task behaved differently from people asked to do it. In the first, show-instructed people chose disambiguating final approaches more often (mean 1.82 against 1.05) 2610.14004 §3.2. In the second, they looped over the coloured tiles more (mean 1.38 revisits per trajectory against 0.10) 2610.14004 §4.2.2.

Maximum-likelihood IRL recovered the reward better from the showing trajectories than from the doing trajectories 2610.14004 §4.3 2610.14004 fig. 2. The distances from the true reward are plotted only, so the size of the gap is not in the text. Card: Showing versus Doing: Teaching by Demonstration.

Adversarial IRL on human keypoints, then cobot transfer

Visual IRL takes RGB-D video of a person and finds 3D arm keypoints and object positions in each frame. The state is the end-effector position, the object position and the object label, and the action is the end-effector displacement between two steps 2412.11360 §IV.A 2412.11360 §V.A. Adversarial IRL (AIRL) fits a discriminator to expert and generated pairs, and the reward is its log-odds 2412.11360 eq. 2 2412.11360 eq. 4. The policy is trained with TRPO 2412.11360 §IV.A. The discriminator architecture, learning rate and discount are not stated 2412.11360 §IV.A.

The reward is only half of the method. A separate step maps the human arm pose to initial cobot joint angles with a learned model, then refines the angles with inverse kinematics so the gripper reaches the target while the joint changes stay small 2412.11360 §IV.B 2412.11360 eq. 6. The paper's aim is motion that looks human, not only motion that reaches the goal 2412.11360 §I.

The paper reports 90.4% learned-behaviour accuracy against 83.3% for a discrete-input baseline, in the text only 2412.11360 §V.D. On one-object onion sorting, the neuro-symbolic model takes 14.75 s against 30.87 s for RRT, with jerkiness of 264.92 against 914.13 degrees 2412.11360 tab. 5. On bottle pouring it takes 33.80 s against 44.12 s for RRT-connect 2412.11360 tab. 6. The pouring trigger is a simple tilt, so the task measures motion, not pouring quality 2412.11360 §V.A. The data come from one lab with 10 trials per task, and the dataset is not released 2412.11360 §V.D. Card: Visual IRL for Human-Like Robotic Manipulation.

Optimal-transport reward from one or two expert videos

TemporalOT learns a policy from one or two expert videos, with no task reward. Each agent step is relabelled with the negative cost it carries under an optimal-transport plan between agent and expert frames 2410.21795 eq. 5. Two changes make the reward aware of order. The cost averages the cosine distance over a short window of consecutive frame pairs 2410.21795 eq. 6, and the plan is restricted to a band around the diagonal 2410.21795 eq. 9. The encoder is a fixed ImageNet ResNet50 2410.21795 §4.1.

On nine Meta-World tasks with two demonstrations and five seeds, the average success rate is 61.1, against 35.9 for ADS and 20.8 for the sparse task reward 2410.21795 tab. 1. These are author-reported numbers, and some standard deviations are large, for example 42.2 (44.5) for ADS on Basketball 2410.21795 tab. 1. When the expert is sampled 3x or 4x faster, success falls on Basketball and Door-open 2410.21795 tab. 3. The method assumes that the agent and the expert move at a similar speed 2410.21795 §3.2.3.

Each trajectory needs one Sinkhorn solve per demonstration, so cost grows with the number of demonstrations 2410.21795 §6. Wall-clock time is one hour for state runs and three for pixel runs on one GPU 2410.21795 app. B.4. The repository linked from the paper has no license file, so this page cites only the paper Card: Robot Policy Learning with Temporal Optimal Transport Reward.

Reward machine inferred from unlabelled demonstrations

This variant infers the task structure as well as the reward. Frames from several demonstrations are embedded and clustered with DBSCAN. The frame nearest each cluster mean becomes a prototype, which gives one state of a reward machine and one proposition, true when a frame is within a distance threshold of the prototype 2412.10096 §4.3 2412.10096 §4.4. Transitions are read from the order of the propositions in each demonstration [arxiv:2412.10096 alg. 1]. A DQN then learns on the inferred machine, with potential-based shaping 2412.10096 §5 2412.10096 eq. 6.

The paper publishes no numeric table. On five simulated Franka Panda block tasks, learning curves of the inferred-machine agent are similar to those of an agent given the ground-truth machine, except on Stack-3, where the baseline plateaus higher 2412.10096 fig. 5. The thresholds and clustering parameters are tuned per task, and the tuning method is not stated 2412.10096 tab. 1. The demonstrations are generated by a scripted expert, and the paper describes them in two different ways 2412.10096 §6. The method is tested only in simulation 2412.10096 §7. Card: Reward Machine Inference for Robotic Manipulation.

Demonstrations in the prompt of a frozen LLM

A frozen language model can continue a sequence of numbers, given a few examples in its prompt, with no training step. "Large Language Models as General Pattern Machines" applies this to robot motion. In the table-sweeping task, 30 demonstrations are binned into numbers, the first two thirds are given as context, and the model predicts the last third 2307.04721 §5 2307.04721 app. B.2. The paper reports that text-davinci-003 has the lowest variance over 30 trials 2307.04721 app. B.2. For the marker task, the prompt holds reward-labelled trajectories, and the model writes new ones expected to score higher 2307.04721 §6.

This route has real limits. Each action is one model query, and the paper names latency, context length and monetary cost as blockers for real systems 2307.04721 §7 2307.04721 §1. The grid and CartPole experiments are capped at 1024 tokens of context 2307.04721 app. C.2 2307.04721 app. C.3. The main results use hosted OpenAI and PaLM models 2307.04721 tab. 1 2307.04721 tab. 2. Most robot results are single runs or small samples 2307.04721 §5 2307.04721 §6. Card: Large Language Models as General Pattern Machines.

Reinforcement learning without demonstrations (for contrast)

The survey also describes a reinforcement learning scheme for deformable objects that does not need initial demonstrations, attributing it to Wu et al. (2019) 2105.01767 §5.1. It is listed here because a reader choosing between LfD and RL needs both options. No comparison of the two is reported in the survey. AnyTeleop's pure RL baseline, which uses no demonstrations, scores below its demonstration-based policy in all six simulated settings 2307.04577 §7.1 2307.04577 tab. 3.

When to use which

The survey gives no thresholds for these choices. The cards test different tasks, robots, simulators and metrics, so their numbers do not rank the variants against each other. A preference between them needs a comparison on your own task.

Implementation sketch

The sketch shows the common core of the trained variants, not any one paper's code. The LLM-prompt variant has no training step and is left out. The pose-matching variant has no policy, so it appears only as a search. Where the variants differ, the comment says so.

# Collect expert demonstrations; teleoperation is one way to record them
# [arxiv:2307.04577 §4] [arxiv:2105.01767 §5.1]
demos = record_expert_runs(task, n=num_demos)

# Keep keyframes only: joint velocity near zero, gripper unchanged
# [arxiv:2209.05451 §3.2]; the threshold is not stated
keyframes = filter_keyframes(demos)

# Planning variant: fit a trajectory and force model, then roll it out
# [arxiv:2105.01767 §5.1]
model = fit_trajectory_and_force(demos)
plan = model.rollout(initial_state)

# Behaviour cloning variant: cross-entropy on keyframe actions
# [arxiv:2209.05451 §3.4]
policy = train_cross_entropy(policy, keyframes)

# Many-task variant: a task embedding conditions the policy, and human
# interventions during rollouts join the data [arxiv:2202.02005 §5.3] [§4]
policy = train_conditioned(policy, demos, task=sentence_encoder(command))

# Lookup variant: fine-tune the encoder on the frames, no action labels,
# then average the actions of the k nearest frames [arxiv:2112.01511 §3.2]
encoder = finetune_byol(encoder, demos.frames)
neighbours = k_nearest(encoder(demos.frames), encoder(query), k=10)
action = softmax(-neighbours.dist) @ demos.actions[neighbours.idx]

# Pose-matching variant: no policy; search for the pose whose query
# descriptors match the averaged demonstration descriptor [arxiv:2112.05124 eq. 11]
target = mean([pose_descriptor(d.pick, d.cloud, Q) for d in demos])
pose = solve_pose(cloud, target, Q)

# RL bootstrap variant: start RL from the demonstrations
# [arxiv:2105.01767 §3.2]
policy = pretrain_on(demos)
policy = reinforce(policy, env, reward)

# Preference variant: fit the reward to pairwise labels on clips, then RL
# on it; the labels must keep coming during training [arxiv:1811.06521 §2.4]
reward = fit_reward_from_preferences(clip_pairs, labels)
policy = reinforce(policy, env, reward)

# Reward recovery variant: infer a reward, then run RL on it
# [arxiv:2105.01767 §3.2] [arxiv:2610.14004 §4.3]
reward = inverse_rl(demos)
policy = reinforce(policy, env, reward)

Contradictions

None known between the cards on any reported number. The cards differ in what a demonstration is. The survey assumes an expert performing the task 2105.01767 §5.1. "Showing versus Doing" argues that a demonstration given to teach carries different information from one given to do, and its experiments support that for reward recovery 2610.14004 §4.3.

Within one card, the reported count of games differs. The Ibarz et al. abstract says the method beats imitation in 7 of 9 Atari games 1811.06521 abstract, while its results section says 8 of 9, excluding only Private Eye 1811.06521 §3.1. The card treats 7 as the conservative figure.

Several results are not comparable across cards. VINN and PerAct test different robots, tasks and metrics 2112.01511 tab. I 2209.05451 tab. 1. Neural Descriptor Fields reports simulated pick-and-place of mugs, bowls and bottles against Dense Object Nets, so its numbers cannot be ranked against either 2112.05124 tab. I. BC-Z's 32% is a held-out success rate on real-robot tasks, and PerAct's 1.33x and 2.83x are gains over a baseline in simulation, so the two cannot be ranked 2202.02005 tab. 3 2209.05451 tab. 1. The VINN paper also reports that frozen-representation behaviour cloning beats VINN in the unmodified scene 2112.01511 tab. II. The newer cards are not comparable with these either: TemporalOT reports success rates in simulation, Visual IRL reports time and jerkiness on real cobots, and the reward-machine paper shows curves only 2410.21795 tab. 1 2412.11360 tab. 5 2412.10096 fig. 5.

TemporalOT's own tables disagree on one number. Table 3 gives 89.0 for Button-press at 1x, while Table 1 gives 92.4 for TemporalOT there and 89.0 for ADS 2410.21795 tab. 3 2410.21795 tab. 1. The card treats the Table 1 value as the TemporalOT result.

Papers

• Reward learning from human preferences and demonstrations in Atari: a reward model fitted to pairwise preferences, combined with expert demonstrations for a DQfD agent on nine Atari games.
• Challenges and Outlook in Robotic Manipulation of Deformable Objects: a survey and position paper on deformable-object manipulation. It describes learning from demonstration as a planning method and as a way to start end-to-end learning, and reports no experiments of its own.
• BC-Z: Zero-Shot Task Generalization with Robotic Imitation Learning: one image-to-action policy over 100 real-robot tasks, conditioned on language or video, with human interventions in the training data.
• The Surprising Effectiveness of Representation Learning for Visual Imitation: VINN, a fine-tuned encoder with a nearest-neighbour action lookup over demonstration frames, with real-robot door-opening results.
• Perceiver-Actor: A Multi-Task Transformer for Robotic Manipulation: keyframe behaviour cloning with a voxel Transformer and language goal, for 18 simulated tasks and 7 real-robot tasks.
• AnyTeleop: A General Vision-Based Dexterous Robot Arm-Hand Teleoperation System: camera-based teleoperation for collecting demonstrations, with an imitation-learning comparison in simulation.
• DexPilot: Vision Based Teleoperation of Dexterous Robotic Hand-Arm System: vision-only teleoperation with four RGB-D cameras and a fingertip-vector retargeting step; no learned policy.
• GEX: Democratizing Dexterity with Fully-Actuated Dexterous Hand and Exoskeleton Glove: a low-cost three-finger hand and exoskeleton glove for teleoperation, with qualitative results only.
• Large Language Models as General Pattern Machines: a frozen LLM continues few-shot sequences, including robot trajectories, with no training.
• DriveMLM: Aligning Multi-Modal Large Language Models with Behavioral Planning States for Autonomous Driving: a multimodal LLM outputs driving decisions for a motion planner, trained on expert CARLA driving with rule-derived labels.
• Information-Theoretic Graph Fusion with Vision-Language-Action Model for Policy Reasoning and Dual Robotic Control: a scene graph from one human RGB-D demonstration, planned by a 7B language model with a separate action head for two arms.
• Showing versus Doing: Teaching by Demonstration: models demonstrations as teaching, and compares reward recovery from showing and doing trajectories.
• Neural Descriptor Fields: SE(3)-Equivariant Object Representations for Manipulation: point descriptors learned without keypoint labels, used to match demonstrated grasp and placement poses; simulated results with 10 demonstrations, and qualitative real-robot examples.
• A Survey of Imitation Learning: Algorithms, Recent Developments, and Challenges: a survey of behavioural cloning, inverse RL, adversarial and observation-based imitation, with open challenges; it reports no experiments.
• Robot Policy Learning with Temporal Optimal Transport Reward: an optimal-transport reward from one or two expert videos, made order-aware with a context window and a temporal band, evaluated on nine Meta-World tasks.
• Reward Machine Inference for Robotic Manipulation: a reward machine inferred from unlabelled block-stacking demonstrations by clustering frame features, used to train a DQN in simulation.
• Visual IRL for Human-Like Robotic Manipulation: AIRL on human keypoints and object positions, with a neuro-symbolic mapping and inverse kinematics to move the learned motion onto a cobot.