Amazon Alexa
Abstract:Large language models (LLMs) encode vast amounts of world knowledge acquired via training on large web-scale datasets crawled from the internet. However, these datasets typically exhibit a geographical bias towards English-speaking Western countries. This results in LLMs producing biased or hallucinated responses to queries that require answers localized to other geographical regions. In this work, we introduce a new benchmark named LoFTI (Localization and Factuality Transfer to Indian Locales) that can be used to evaluate an LLM's localization and factual text transfer capabilities. LoFTI consists of factual statements about entities in source and target locations; the source locations are spread across the globe and the target locations are all within India with varying degrees of hyperlocality (country, states, cities). The entities span a wide variety of categories. We use LoFTI to evaluate Mixtral, GPT-4 and two other Mixtral-based approaches well-suited to the task of localized factual transfer. We demonstrate that LoFTI is a high-quality evaluation benchmark and all the models, including GPT-4, produce skewed results across varying levels of hyperlocality.
Abstract:Human language has been described as a system that makes \textit{use of finite means to express an unlimited array of thoughts}. Of particular interest is the aspect of compositionality, whereby, the meaning of a compound language expression can be deduced from the meaning of its constituent parts. If artificial agents can develop compositional communication protocols akin to human language, they can be made to seamlessly generalize to unseen combinations. Studies have recognized the role of curiosity in enabling linguistic development in children. In this paper, we seek to use this intrinsic feedback in inducing a systematic and unambiguous protolanguage. We demonstrate how compositionality can enable agents to not only interact with unseen objects but also transfer skills from one task to another in a zero-shot setting: \textit{Can an agent, trained to `pull' and `push twice', `pull twice'?}.
Abstract:Human language has been described as a system that makes use of finite means to express an unlimited array of thoughts. Of particular interest is the aspect of compositionality, whereby, the meaning of a complex, compound language expression can be deduced from the meaning of its constituent parts. If artificial agents can develop compositional communication protocols akin to human language, they can be made to seamlessly generalize to unseen combinations. However, the real question is, how do we induce compositionality in emergent communication? Studies have recognized the role of curiosity in enabling linguistic development in children. It is this same intrinsic urge that drives us to master complex tasks with decreasing amounts of explicit reward. In this paper, we seek to use this intrinsic feedback in inducing a systematic and unambiguous protolanguage in artificial agents. We show in our experiments, how these rewards can be leveraged in training agents to induce compositionality in absence of any external feedback. Additionally, we introduce Comm-gSCAN, a platform for investigating grounded language acquisition in 2D-grid environments. Using this, we demonstrate how compositionality can enable agents to not only interact with unseen objects, but also transfer skills from one task to other in zero-shot (Can an agent, trained to pull and push twice, pull twice?)