While pioneering deep learning methods have made great strides in analyzing electronic health record (EHR) data, they often struggle to fully capture the semantics of diverse medical codes from limited data. The integration of external knowledge from Large Language Models (LLMs) presents a promising avenue for improving healthcare predictions. However, LLM analyses may exhibit significant variance due to ambiguity problems and inconsistency issues, hindering their effective utilization. To address these challenges, we propose IntelliCare, a novel framework that leverages LLMs to provide high-quality patient-level external knowledge and enhance existing EHR models. Concretely, IntelliCare identifies patient cohorts and employs task-relevant statistical information to augment LLM understanding and generation, effectively mitigating the ambiguity problem. Additionally, it refines LLM-derived knowledge through a hybrid approach, generating multiple analyses and calibrating them using both the EHR model and perplexity measures. Experimental evaluations on three clinical prediction tasks across two large-scale EHR datasets demonstrate that IntelliCare delivers significant performance improvements to existing methods, highlighting its potential in advancing personalized healthcare predictions and decision support systems.

Code benchmarks such as HumanEval are widely adopted to evaluate Large Language Models' (LLMs) coding capabilities. However, there is an unignorable programming language bias in existing code benchmarks -- over 95% code generation benchmarks are dominated by Python, leaving the LLMs' capabilities in other programming languages such as Java and C/C++ unknown. Moreover, coding task bias is also crucial. Most benchmarks focus on code generation capability, while benchmarks for code reasoning (given input, reasoning output; and given output, reasoning input), an essential coding capability, are insufficient. Yet, constructing multi-lingual benchmarks can be expensive and labor-intensive, and codes in contest websites such as Leetcode suffer from data contamination during training. To fill this gap, we propose CRUXEVAL-X, a multi-lingual code reasoning benchmark that contains 19 programming languages. It comprises at least 600 subjects for each language, along with 19K content-consistent tests in total. In particular, the construction pipeline of CRUXEVAL-X works in a fully automated and test-guided manner, which iteratively generates and repairs based on execution feedback. Also, to cross language barriers (e.g., dynamic/static type systems in Python/C++), we formulated various transition rules between language pairs to facilitate translation. Our intensive evaluation of 24 representative LLMs reveals the correlation between language pairs. For example, TypeScript and JavaScript show a significant positive correlation, while Racket has less correlation with other languages. More interestingly, even a model trained solely on Python can achieve at most 34.4% Pass@1 in other languages, revealing the cross-language generalization of LLMs.

In the era of large language models (LLMs), the task of ``System I''~-~the fast, unconscious, and intuitive tasks, e.g., sentiment analysis, text classification, etc., have been argued to be successfully solved. However, sarcasm, as a subtle linguistic phenomenon, often employs rhetorical devices like hyperbole and figuration to convey true sentiments and intentions, involving a higher level of abstraction than sentiment analysis. There is growing concern that the argument about LLMs' success may not be fully tenable when considering sarcasm understanding. To address this question, we select eleven SOTA LLMs and eight SOTA pre-trained language models (PLMs) and present comprehensive evaluations on six widely used benchmark datasets through different prompting approaches, i.e., zero-shot input/output (IO) prompting, few-shot IO prompting, chain of thought (CoT) prompting. Our results highlight three key findings: (1) current LLMs underperform supervised PLMs based sarcasm detection baselines across six sarcasm benchmarks. This suggests that significant efforts are still required to improve LLMs' understanding of human sarcasm. (2) GPT-4 consistently and significantly outperforms other LLMs across various prompting methods, with an average improvement of 14.0\%$\uparrow$. Claude 3 and ChatGPT demonstrate the next best performance after GPT-4. (3) Few-shot IO prompting method outperforms the other two methods: zero-shot IO and few-shot CoT. The reason is that sarcasm detection, being a holistic, intuitive, and non-rational cognitive process, is argued not to adhere to step-by-step logical reasoning, making CoT less effective in understanding sarcasm compared to its effectiveness in mathematical reasoning tasks.

Google DeepMind workers have signed a letter calling on the company to drop contracts with military organizations, according to a report by Time. The document was drafted on May 16 of this year. Around 200 people signed the document, which amounts to five percent of the total headcount of DeepMind. For the uninitiated, DeepMind is one of Google's AI divisions and the letter states that adopting military contracts runs afoul of the company's own AI rules. The letter was sent out as internal concerns began circulating within the AI lab that the tech was allegedly being sold to military organizations via cloud contracts.

Ayah Bdeir, a senior advisor to Mozilla and a participant in OSI's process, says certain parts of the open-source definition were relatively easy to agree upon, including the need to reveal model weights (the parameters that help determine how an AI model generates an output). Other parts of the deliberations were more contentious, particularly the question of how public training data should be. The lack of transparency about where training data comes from has led to innumerable lawsuits against big AI companies, from makers of large language models like OpenAI to music generators like Suno, which do not disclose much about their training sets beyond saying they contain "publicly accessible information." Ultimately, the new definition requires that open-source models provide information about the training data to the extent that "a skilled person can recreate a substantially equivalent system using the same or similar data." It's not a blanket requirement to share all training data sets, but it also goes further than what many proprietary models or even ostensibly open-source models do today. "Insisting on an ideologically pristine kind of gold standard that actually will not effectively be met by anybody ends up backfiring," Bdeir says.

Nearly two years ago, ChatGPT's AI writing powers set off a firestorm in classrooms. How would teachers be able to determine which assignments were actually authored by the student? A host of AI-powered services answered the call. Today, there are even more services promising to catch AI cheaters. "My hand cramped up so much," my eldest son complained about his AP World History course he took last year, and the requirement to handwrite all papers and tests because of AI concerns.

Nearly 200 workers inside Google DeepMind, the company's AI division, signed a letter calling on the tech giant to drop its contracts with military organizations earlier this year, according to a copy of the document reviewed by TIME and five people with knowledge of the matter. The letter circulated amid growing concerns inside the AI lab that its technology is being sold to militaries engaged in warfare, in what the workers say is a violation of Google's own AI rules. The letter is a sign of a growing dispute within Google between at least some workers in its AI division--which has pledged to never work on military technology--and its Cloud business, which has contracts to sell Google services, including AI developed inside DeepMind, to several governments and militaries including those of Israel and the United States. The signatures represent some 5% of DeepMind's overall headcount--a small portion to be sure, but a significant level of worker unease for an industry where top machine learning talent is in high demand. The DeepMind letter, dated May 16 of this year, begins by stating that workers are "concerned by recent reports of Google's contracts with military organizations."

In this paper, we demonstrate that non-generative, small-sized models such as FinBERT and FinDRoBERTa, when fine-tuned, can outperform GPT-3.5 and GPT-4 models in zero-shot learning settings in sentiment analysis for financial news. These fine-tuned models show comparable results to GPT-3.5 when it is fine-tuned on the task of determining market sentiment from daily financial news summaries sourced from Bloomberg. To fine-tune and compare these models, we created a novel database, which assigns a market score to each piece of news without human interpretation bias, systematically identifying the mentioned companies and analyzing whether their stocks have gone up, down, or remained neutral. Furthermore, the paper shows that the assumptions of Condorcet's Jury Theorem do not hold suggesting that fine-tuned small models are not independent of the fine-tuned GPT models, indicating behavioural similarities. Lastly, the resulted fine-tuned models are made publicly available on HuggingFace, providing a resource for further research in financial sentiment analysis and text classification.

This paper describes the architecture of our system developed for Task 3 of SemEval-2024: Multimodal Emotion-Cause Analysis in Conversations. Our project targets the challenges of subtask 2, dedicated to Multimodal Emotion-Cause Pair Extraction with Emotion Category (MECPE-Cat), and constructs a dual-component system tailored to the unique challenges of this task. We divide the task into two subtasks: emotion recognition in conversation (ERC) and emotion-cause pair extraction (ECPE). To address these subtasks, we capitalize on the abilities of Large Language Models (LLMs), which have consistently demonstrated state-of-the-art performance across various natural language processing tasks and domains. Most importantly, we design an approach of emotion-cause-aware instruction-tuning for LLMs, to enhance the perception of the emotions with their corresponding causal rationales. Our method enables us to adeptly navigate the complexities of MECPE-Cat, achieving a weighted average 34.71% F1 score of the task, and securing the 2nd rank on the leaderboard. The code and metadata to reproduce our experiments are all made publicly available.

Large Language Models (LLMs) have demonstrated exceptional capabilities across various machine learning (ML) tasks. Given the high costs of creating annotated datasets for supervised learning, LLMs offer a valuable alternative by enabling effective few-shot in-context learning. However, these models can produce hallucinations, particularly in domains with incomplete knowledge. Additionally, current methods for knowledge distillation using LLMs often struggle to enhance the effectiveness of both teacher and student models. To address these challenges, we introduce DualChecker, an innovative framework designed to mitigate hallucinations and improve the performance of both teacher and student models during knowledge distillation. DualChecker employs ContextAligner to ensure that the context provided by teacher models aligns with human labeling standards. It also features a dynamic checker system that enhances model interaction: one component re-prompts teacher models with more detailed content when they show low confidence, and another identifies borderline cases from student models to refine the teaching templates. This interactive process promotes continuous improvement and effective knowledge transfer between the models. We evaluate DualChecker using a green innovation textual dataset that includes binary, multiclass, and token classification tasks. The experimental results show that DualChecker significantly outperforms existing state-of-the-art methods, achieving up to a 17% improvement in F1 score for teacher models and 10% for student models. Notably, student models fine-tuned with LLM predictions perform comparably to those fine-tuned with actual data, even in a challenging domain. We make all datasets, models, and code from this research publicly available.