Computation and Language (cs.CL)

Abstract: Reinforcement learning from human feedback (RLHF) is effective at aligning large language models (LLMs) to human preferences, but gathering high quality human preference labels is a key bottleneck. We conduct a head-to-head comparison of RLHF vs. RL from AI Feedback (RLAIF) - a technique where preferences are labeled by an off-the-shelf LLM in lieu of humans, and we find that they result in similar improvements. On the task of summarization, human evaluators prefer generations from both RLAIF and RLHF over a baseline supervised fine-tuned model in ~70% of cases. Furthermore, when asked to rate RLAIF vs. RLHF summaries, humans prefer both at equal rates. These results suggest that RLAIF can yield human-level performance, offering a potential solution to the scalability limitations of RLHF.

Abstract: Topic modeling and text mining are subsets of Natural Language Processing with relevance for conducting meta-analysis (MA) and systematic review (SR). For evidence synthesis, the above NLP methods are conventionally used for topic-specific literature searches or extracting values from reports to automate essential phases of SR and MA. Instead, this work proposes a comparative topic modeling approach to analyze reports of contradictory results on the same general research question. Specifically, the objective is to find topics exhibiting distinct associations with significant results for an outcome of interest by ranking them according to their proportional occurrence and consistency of distribution across reports of significant results. The proposed method was tested on broad-scope studies addressing whether supplemental nutritional compounds significantly benefit macular degeneration (MD). Eight compounds were identified as having a particular association with reports of significant results for benefitting MD. Six of these were further supported in terms of effectiveness upon conducting a follow-up literature search for validation (omega-3 fatty acids, copper, zeaxanthin, lutein, zinc, and nitrates). The two not supported by the follow-up literature search (niacin and molybdenum) also had the lowest scores under the proposed methods ranking system, suggesting that the proposed method's score for a given topic is a viable proxy for its degree of association with the outcome of interest. These results underpin the proposed methods potential to add specificity in understanding effects from broad-scope reports, elucidate topics of interest for future research, and guide evidence synthesis in a systematic and scalable way.

Abstract: Shannon, in his seminal paper introducing information theory, divided the communication into three levels: technical, semantic, and effectivenss. While the technical level is concerned with accurate reconstruction of transmitted symbols, the semantic and effectiveness levels deal with the inferred meaning and its effect on the receiver. Thanks to telecommunications, the first level problem has produced great advances like the internet. Large Language Models (LLMs) make some progress towards the second goal, but the third level still remains largely untouched. The third problem deals with predicting and optimizing communication for desired receiver behavior. LLMs, while showing wide generalization capabilities across a wide range of tasks, are unable to solve for this. One reason for the underperformance could be a lack of "behavior tokens" in LLMs' training corpora. Behavior tokens define receiver behavior over a communication, such as shares, likes, clicks, purchases, retweets, etc. While preprocessing data for LLM training, behavior tokens are often removed from the corpora as noise. Therefore, in this paper, we make some initial progress towards reintroducing behavior tokens in LLM training. The trained models, other than showing similar performance to LLMs on content understanding tasks, show generalization capabilities on behavior simulation, content simulation, behavior understanding, and behavior domain adaptation. Using a wide range of tasks on two corpora, we show results on all these capabilities. We call these models Large Content and Behavior Models (LCBMs). Further, to spur more research on LCBMs, we release our new Content Behavior Corpus (CBC), a repository containing communicator, message, and corresponding receiver behavior.

Abstract: The capabilities and use cases of automatic natural language processing (NLP) have grown significantly over the last few years. While much work has been devoted to understanding how humans deal with discourse connectives, this phenomenon is understudied in computational systems. Therefore, it is important to put NLP models under the microscope and examine whether they can adequately comprehend, process, and reason within the complexity of natural language. In this chapter, we introduce the main mechanisms behind automatic sentence processing systems step by step and then focus on evaluating discourse connective processing. We assess nine popular systems in their ability to understand English discourse connectives and analyze how context and language understanding tasks affect their connective comprehension. The results show that NLP systems do not process all discourse connectives equally well and that the computational processing complexity of different connective kinds is not always consistently in line with the presumed complexity order found in human processing. In addition, while humans are more inclined to be influenced during the reading procedure but not necessarily in the final comprehension performance, discourse connectives have a significant impact on the final accuracy of NLP systems. The richer knowledge of connectives a system learns, the more negative effect inappropriate connectives have on it. This suggests that the correct explicitation of discourse connectives is important for computational natural language processing.

Abstract: Marketers spend billions of dollars on advertisements but to what end? At the purchase time, if customers cannot recognize a brand for which they saw an ad, the money spent on the ad is essentially wasted. Despite its importance in marketing, until now, there has been no study on the memorability of ads in the ML literature. Most studies have been conducted on short-term recall (<5 mins) on specific content types like object and action videos. On the other hand, the advertising industry only cares about long-term memorability (a few hours or longer), and advertisements are almost always highly multimodal, depicting a story through its different modalities (text, images, and videos). With this motivation, we conduct the first large scale memorability study consisting of 1203 participants and 2205 ads covering 276 brands. Running statistical tests over different participant subpopulations and ad-types, we find many interesting insights into what makes an ad memorable - both content and human factors. For example, we find that brands which use commercials with fast moving scenes are more memorable than those with slower scenes (p=8e-10) and that people who use ad-blockers remember lower number of ads than those who don't (p=5e-3). Further, with the motivation of simulating the memorability of marketing materials for a particular audience, ultimately helping create one, we present a novel model, Sharingan, trained to leverage real-world knowledge of LLMs and visual knowledge of visual encoders to predict the memorability of a content. We test our model on all the prominent memorability datasets in literature (both images and videos) and achieve state of the art across all of them. We conduct extensive ablation studies across memory types, modality, brand, and architectural choices to find insights into what drives memory.

Abstract: Many LLMs are trained to perform zero-shot or few-shot inference using instruction-based prompts. Crafting prompts for these LLMs typically requires the user to provide a detailed task description, examples of context and completion, and single example of context for inference. This regular prompt baseline is referred to as SinglePrompt in this paper. However, for NLP tasks where each data point for inference is not necessarily lengthy, the token count for instructions and few-shot examples in the prompt may be considerably larger than that of the data point, resulting in lower token-resource utilization compared with encoder-based models like fine-tuned BERT. This cost-efficiency issue, affecting inference speed and compute budget, counteracts the many benefits LLMs have to offer. This paper aims to alleviate the preceding problem by batching multiple data points into a single prompt, a prompting strategy we refer to as BatchPrompt. This strategy increases the density of data points, which in turn leads to improved token utilization. Applying BatchPrompt naively, however, is very challenging due to significant performance degradation, as observed in our experiments. We also noticed varying inference outcomes for the same data point appearing in different positions within a prompt. To address the quality issue while remain high token-resource utilization, we introduce Batch Permutation and Ensembling for BatchPrompt, a simple way that recovers labeling quality through majority votes from data points placed in varying positions in a batch at the price of more token usage. To counterbalance the additional token usage caused by the voting process, we further propose Self-reflection-guided EArly Stopping, which can terminate the voting process early for data points the LLM confidently handles.