State-of-the-art methods for Human-AI Teaming and Zero-shot Cooperation focus on task completion i.e. task rewards, as the sole evaluation metric while being agnostic to `how' the two agents work with each other. Furthermore, subjective user studies only offer limited insight into the quality of cooperation existing within the team. Specifically, we are interested in understanding the cooperative behaviors arising within the team when trained agents are paired with humans - a problem that has been overlooked by the existing literature. To formally address this problem, we propose the concept of constructive interdependence - measuring how much agents rely on each other’s actions to achieve the shared goal - as a key metric for evaluating cooperation in human-agent teams. We measure interdependence in terms of action interactions in a STRIPS formalism, and define metrics that allow us to assess the degree of reliance between the agents' actions. We pair state-of-the-art agents with learned human models as well as human participants in a user study for the popular Overcooked domain, and evaluate the task reward and teaming performance for these human-agent teams. While prior work has claimed that state-of-the-art agents exhibit cooperative behavior based on their high task rewards, our results reveal that these agents often fail to induce cooperation, as evidenced by consistently low interdependence across teams. Furthermore, our analysis reveals that teaming performance is not necessarily correlated with task reward, highlighting that task reward alone cannot reliably measure cooperation arising in a human-agent team.

Recent advances in reasoning-oriented Large Language Models (LLMs) have been driven by the introduction of Chain-of-Thought (CoT) traces, where models generate intermediate reasoning traces before producing an answer.These traces, as in DeepSeek R1, are not only used to guide model inference but also serve as supervision signals for Knowledge Distillation (KD) to improve smaller models.A prevailing but under-examined implicit assumption is that these CoT traces when emitted at inference time are both semantically correct and interpretable for the end-users.While there are reasons to believe that these intermediate tokens help improve solution accuracy, in this work, we question their validity (semantic correctness) and interpretability to the end user.To isolate the effect of trace semantics, we design experiments in the Question Answering (QA) domain using a rule-based problem decomposition method.This enables us to create Supervised Fine-Tuning (SFT) datasets for LLMs where -each QA problem is paired with either verifiably correct or incorrect CoT traces, while always providing the correct final solution.Trace correctness at inference time is then evaluated by checking the accuracy of every sub-step in decomposed reasoning chains.To assess end-user interpretability, we finetune LLMs with three additional types of CoT traces: R1 traces, R1 trace summaries, and post-hoc explanations of R1 traces.We further conduct a human-subject study with 100 participants asking them to rate the interpretability of each trace type on a standardized Likert scale.Our experiments reveal two key findings -(1) CoT trace correctness is not reliably correlated with the model's generation of correct final answers: correct traces led to correct solutions only for 28% test-set problems while incorrect traces don't necessarily degrade solution accuracy.(2) In end-user interpretability studies, fine-tuning * Work done while a PhD student at ASU, currently at Samsung Research America on verbose R1 traces prooduced the best model performance but these traces were rated as least interpretable by users, scoring on average 3.39 for interpretability and 4.59 for cognitive load metrics on a 5-point Likert scale.In contrast, the decomposed traces that are judged significantly more interpretable don't lead to comparable solution accuracy.Together, these findings challenge the assumption in question suggesting that researchers and practitioners should decouple model supervision objectives from end-user-facing trace design. 1

We argue that auto-regressive LLMs cannot, by themselves, do planning or self-verification (which is after all a form of reasoning), and shed some light on the reasons for misunderstandings in the literature. We will also argue that LLMs should be viewed as universal approximate knowledge sources that have much more meaningful roles to play in planning/reasoning tasks beyond simple front-end/back-end format translators. We present a vision of LLM-Modulo Frameworks that combine the strengths of LLMs with external model-based verifiers in a tighter bi-directional interaction regime. We will show how the models driving the external verifiers themselves can be acquired with the help of LLMs. We will also argue that rather than simply pipelining LLMs and symbolic components, this LLM-Modulo Framework provides a better neuro-symbolic approach that offers tighter integration between LLMs and symbolic components, and allows extending the scope of model-based planning/reasoning regimes towards more flexible knowledge, problem and preference specifications.

In this paper, we consider an important class of Partially Observable Markov Decision Processes (POMDP) with unknown parameters, which contains the Wordle puzzle as a special case. For this class of POMDP, we develop a new on-line solution method, which is based on the rollout approach. Our method relies on the use of a base heuristic policy and guarantees cost improvement over that policy. When applied to Wordle, our algorithm solves the puzzle on-line. The performance is within 0.4% of the known optimal results, and is substantially better than that of the base heuristic policies we have tested.

Conventional works that learn grasping affordance from demonstrations need to explicitly predict grasping configurations, such as gripper approaching angles or grasping preshapes. Classic motion planners could then sample trajectories by using such predicted configurations. In this work, our goal is instead to fill the gap between affordance discovery and affordance-based policy learning by integrating the two objectives in an end-to-end imitation learning framework based on deep neural networks. From a psychological perspective, there is a close association between attention and affordance. Therefore, with an end-to-end neural network, we propose to learn affordance cues as visual attention that serves as a useful indicating signal of how a demonstrator accomplishes tasks, instead of explicitly modeling affordances. To achieve this, we propose a contrastive learning framework that consists of a Siamese encoder and a trajectory decoder. We further introduce a coupled triplet loss to encourage the discovered affordance cues to be more affordance-relevant. Our experimental results demonstrate that our model with the coupled triplet loss achieves the highest grasping success rate in a simulated robot environment. Our project website can be accessed at1.