While the situation has improved for text-only models, it again seems to be the case currently that multimodal (text and image) models develop faster than ways to evaluate them. In this paper, we bring a recently developed evaluation paradigm from text models to multimodal models, namely evaluation through the goal-oriented game (self) play, complementing reference-based and preference-based evaluation. Specifically, we define games that challenge a model's capability to represent a situation from visual information and align such representations through dialogue. We find that the largest closed models perform rather well on the games that we define, while even the best open-weight models struggle with them. On further analysis, we find that the exceptional deep captioning capabilities of the largest models drive some of the performance. There is still room to grow for both kinds of models, ensuring the continued relevance of the benchmark.

Incremental models that process sentences one token at a time will sometimes encounter points where more than one interpretation is possible. Causal models are forced to output one interpretation and continue, whereas models that can revise may edit their previous output as the ambiguity is resolved. In this work, we look at how restart-incremental Transformers build and update internal states, in an effort to shed light on what processes cause revisions not viable in autoregressive models. We propose an interpretable way to analyse the incremental states, showing that their sequential structure encodes information on the garden path effect and its resolution. Our method brings insights on various bidirectional encoders for contextualised meaning representation and dependency parsing, contributing to show their advantage over causal models when it comes to revisions.

It has been established in recent work that Large Language Models (LLMs) can be prompted to "self-play" conversational games that probe certain capabilities (general instruction following, strategic goal orientation, language understanding abilities), where the resulting interactive game play can be automatically scored. In this paper, we take one of the proposed frameworks for setting up such game-play environments, and further test its usefulness as an evaluation instrument, along a number of dimensions: We show that it can easily keep up with new developments while avoiding data contamination, we show that the tests implemented within it are not yet saturated (human performance is substantially higher than that of even the best models), and we show that it lends itself to investigating additional questions, such as the impact of the prompting language on performance. We believe that the approach forms a good basis for making decisions on model choice for building applied interactive systems, and perhaps ultimately setting up a closed-loop development environment of system and simulated evaluator.

Active participation in a conversation is key to building common ground, since understanding is jointly tailored by producers and recipients. Overhearers are deprived of the privilege of performing grounding acts and can only conjecture about intended meanings. Still, data generation and annotation, modelling, training and evaluation of NLP dialogue models place reliance on the overhearing paradigm. How much of the underlying grounding processes are thereby forfeited? As we show, there is evidence pointing to the impossibility of properly modelling human meta-communicative acts with data-driven learning models. In this paper, we discuss this issue and provide a preliminary analysis on the variability of human decisions for requesting clarification. Most importantly, we wish to bring this topic back to the community's table, encouraging discussion on the consequences of having models designed to only "listen in".

Incremental processing allows interactive systems to respond based on partial inputs, which is a desirable property e.g. in dialogue agents. The currently popular Transformer architecture inherently processes sequences as a whole, abstracting away the notion of time. Recent work attempts to apply Transformers incrementally via restart-incrementality by repeatedly feeding, to an unchanged model, increasingly longer input prefixes to produce partial outputs. However, this approach is computationally costly and does not scale efficiently for long sequences. In parallel, we witness efforts to make Transformers more efficient, e.g. the Linear Transformer (LT) with a recurrence mechanism. In this work, we examine the feasibility of LT for incremental NLU in English. Our results show that the recurrent LT model has better incremental performance and faster inference speed compared to the standard Transformer and LT with restart-incrementality, at the cost of part of the non-incremental (full sequence) quality. We show that the performance drop can be mitigated by training the model to wait for right context before committing to an output and that training with input prefixes is beneficial for delivering correct partial outputs.

While humans process language incrementally, the best language encoders currently used in NLP do not. Both bidirectional LSTMs and Transformers assume that the sequence that is to be encoded is available in full, to be processed either forwards and backwards (BiLSTMs) or as a whole (Transformers). We investigate how they behave under incremental interfaces, when partial output must be provided based on partial input seen up to a certain time step, which may happen in interactive systems. We test five models on various NLU datasets and compare their performance using three incremental evaluation metrics. The results support the possibility of using bidirectional encoders in incremental mode while retaining most of their non-incremental quality. The "omni-directional" BERT model, which achieves better non-incremental performance, is impacted more by the incremental access. This can be alleviated by adapting the training regime (truncated training), or the testing procedure, by delaying the output until some right context is available or by incorporating hypothetical right contexts generated by a language model like GPT-2.

In collaborative goal-oriented settings, the participants are not only interested in achieving a successful outcome, but do also implicitly negotiate the effort they put into the interaction (by adapting to each other). In this work, we propose a challenging interactive reference game that requires two players to coordinate on vision and language observations. The learning signal in this game is a score (given after playing) that takes into account the achieved goal and the players' assumed efforts during the interaction. We show that a standard Proximal Policy Optimization (PPO) setup achieves a high success rate when bootstrapped with heuristic partner behaviors that implement insights from the analysis of human-human interactions. And we find that a pairing of neural partners indeed reduces the measured joint effort when playing together repeatedly. However, we observe that in comparison to a reasonable heuristic pairing there is still room for improvement -- which invites further research in the direction of cost-sharing in collaborative interactions.

Albrecht and Stone (2018) state that modeling of changing behaviors remains an open problem "due to the essentially unconstrained nature of what other agents may do". In this work we evaluate the adaptability of neural artificial agents towards assumed partner behaviors in a collaborative reference game. In this game success is achieved when a knowledgeable Guide can verbally lead a Follower to the selection of a specific puzzle piece among several distractors. We frame this language grounding and coordination task as a reinforcement learning problem and measure to which extent a common reinforcement training algorithm (PPO) is able to produce neural agents (the Guides) that perform well with various heuristic Follower behaviors that vary along the dimensions of confidence and autonomy. We experiment with a learning signal that in addition to the goal condition also respects an assumed communicative effort. Our results indicate that this novel ingredient leads to communicative strategies that are less verbose (staying silent in some of the steps) and that with respect to that the Guide's strategies indeed adapt to the partner's level of confidence and autonomy. Figure 1: An exemplary interaction between a Guide and a Follower that controls the gripper (the black dot).

Clarification requests are a mechanism to help solve communication problems, e.g. due to ambiguity or underspecification, in instruction-following interactions. Despite their importance, even skilful models struggle with producing or interpreting such repair acts. In this work, we test three hypotheses concerning the effects of action taking as an auxiliary task in modelling iCR policies. Contrary to initial expectations, we conclude that its contribution to learning an iCR policy is limited, but some information can still be extracted from prediction uncertainty. We present further evidence that even well-motivated, Transformer-based models fail to learn good policies for when to ask Instruction CRs (iCRs), while the task of determining what to ask about can be more successfully modelled. Considering the implications of these findings, we further discuss the shortcomings of the data-driven paradigm for learning meta-communication acts.

Recent work has proposed a methodology for the systematic evaluation of "Situated Language Understanding Agents"-agents that operate in rich linguistic and non-linguistic contexts-through testing them in carefully constructed interactive settings. Other recent work has argued that Large Language Models (LLMs), if suitably set up, can be understood as (simulators of) such agents. A connection suggests itself, which this paper explores: Can LLMs be evaluated meaningfully by exposing them to constrained game-like settings that are built to challenge specific capabilities? As a proof of concept, this paper investigates five interaction settings, showing that current chat-optimised LLMs are, to an extent, capable to follow game-play instructions. Both this capability and the quality of the game play, measured by how well the objectives of the different games are met, follows the development cycle, with newer models performing better. The metrics even for the comparatively simple example games are far from being saturated, suggesting that the proposed instrument will remain to have diagnostic value. Our general framework for implementing and evaluating games with LLMs is available at https://github.com/clembench .