Recently, there has been a growing demand for conversational speech synthesis (CSS) that generates more natural speech by considering the conversational context. To address this, we introduce JELLY, a novel CSS framework that integrates emotion recognition and context reasoning for generating appropriate speech in conversation by fine-tuning a large language model (LLM) with multiple partial LoRA modules. We propose an Emotion-aware Q-former encoder, which enables the LLM to perceive emotions in speech. The encoder is trained to align speech emotions with text, utilizing datasets of emotional speech. The entire model is then fine-tuned with conversational speech data to infer emotional context for generating emotionally appropriate speech in conversation. Our experimental results demonstrate that JELLY excels in emotional context modeling, synthesizing speech that naturally aligns with conversation, while mitigating the scarcity of emotional conversational speech datasets.

Bearing in mind the limited parametric knowledge of Large Language Models (LLMs), retrieval-augmented generation (RAG) which supplies them with the relevant external knowledge has served as an approach to mitigate the issue of hallucinations to a certain extent. However, uniformly retrieving supporting context makes response generation source-inefficient, as triggering the retriever is not always necessary, or even inaccurate, when a model gets distracted by noisy retrieved content and produces an unhelpful answer. Motivated by these issues, we introduce Semantic Uncertainty Guided Adaptive Retrieval (SUGAR), where we leverage context-based entropy to actively decide whether to retrieve and to further determine between single-step and multi-step retrieval. Our empirical results show that selective retrieval guided by semantic uncertainty estimation improves the performance across diverse question answering tasks, as well as achieves a more efficient inference.

In the constantly changing field of data-driven decision making, accurately predicting future events is crucial for strategic planning in various sectors. The emergence of Large Language Models (LLMs) marks a significant advancement in this area, offering advanced tools that utilise extensive text data for prediction. In this industry paper, we introduce a novel method for AI-driven foresight using LLMs. Building on top of previous research, we employ data on current trends and their trajectories for generating forecasts on 15 different topics. Subsequently, we estimate their probabilities via a multi-step approach based on log probabilities. We show we achieve a Brier score of 0.186, meaning a +26% improvement over random chance and a +19% improvement over widely-available AI systems.

This paper examines the specific obstacles of constructing Retrieval-Augmented Generation(RAG) systems in low-resource languages, with a focus on Persian's complicated morphology and versatile syntax. The research aims to improve retrieval and generation accuracy by introducing Persian-specific models, namely MatinaRoberta(a masked language model) and MatinaSRoberta(a fine-tuned Sentence-BERT), along with a comprehensive benchmarking framework. Three datasets-general knowledge(PQuad), scientifically specialized texts, and organizational reports, were used to assess these models after they were trained on a varied corpus of 73.11 billion Persian tokens. The methodology involved extensive pretraining, fine-tuning with tailored loss functions, and systematic evaluations using both traditional metrics and the Retrieval-Augmented Generation Assessment framework. The results show that MatinaSRoberta outperformed previous embeddings, achieving superior contextual relevance and retrieval accuracy across datasets. Temperature tweaking, chunk size modifications, and document summary indexing were explored to enhance RAG setups. Larger models like Llama-3.1 (70B) consistently demonstrated the highest generation accuracy, while smaller models faced challenges with domain-specific and formal contexts. The findings underscore the potential for developing RAG systems in Persian through customized embeddings and retrieval-generation settings and highlight the enhancement of NLP applications such as search engines and legal document analysis in low-resource languages.

Malware analysis is a complex process of examining and evaluating malicious software's functionality, origin, and potential impact. This arduous process typically involves dissecting the software to understand its components, infection vector, propagation mechanism, and payload. Over the years, deep reverse engineering of malware has become increasingly tedious, mainly due to modern malicious codebases' fast evolution and sophistication. Essentially, analysts are tasked with identifying the elusive needle in the haystack within the complexities of zero-day malware, all while under tight time constraints. Thus, in this paper, we explore leveraging Large Language Models (LLMs) for semantic malware analysis to expedite the analysis of known and novel samples. Built on GPT-4o-mini model, \msp is designed to augment malware analysis for Android through a hierarchical-tiered summarization chain and strategic prompt engineering. Additionally, \msp performs malware categorization, distinguishing potential malware from benign applications, thereby saving time during the malware reverse engineering process. Despite not being fine-tuned for Android malware analysis, we demonstrate that through optimized and advanced prompt engineering \msp can achieve up to 77% classification accuracy while providing highly robust summaries at functional, class, and package levels. In addition, leveraging the backward tracing of the summaries from package to function levels allowed us to pinpoint the precise code snippets responsible for malicious behavior.

Code Large Language Models (LLMs) demonstrate great versatility in adapting to various downstream tasks, including code generation and completion, as well as bug detection and fixing. However, Code LLMs often fail to capture existing coding standards, leading to the generation of code that conflicts with the required design patterns for a given project. As a result, developers must post-process to adapt the generated code to the project's design norms. In this work, we empirically investigate the biases of Code LLMs in software development. Through carefully designed experiments, we assess the models' understanding of design patterns across recognition, comprehension, and generation. Our findings reveal that biases in Code LLMs significantly affect the reliability of downstream tasks.

Conventional recommendation systems (RSs) are typically optimized to enhance performance metrics uniformly across all training samples, inadvertently overlooking the needs of diverse user populations. The performance disparity among various populations can harm the model's robustness to sub-populations due to the varying user properties. While large language models (LLMs) show promise in enhancing RS performance, their practical applicability is hindered by high costs, inference latency, and degraded performance on long user queries. To address these challenges, we propose a hybrid task allocation framework designed to promote social good by equitably serving all user groups. By adopting a two-phase approach, we promote a strategic assignment of tasks for efficient and responsible adaptation of LLMs. Our strategy works by first identifying the weak and inactive users that receive a suboptimal ranking performance by RSs. Next, we use an in-context learning approach for such users, wherein each user interaction history is contextualized as a distinct ranking task. We evaluate our hybrid framework by incorporating eight different recommendation algorithms and three different LLMs -- both open and close-sourced. Our results on three real-world datasets show a significant reduction in weak users and improved robustness to subpopulations without disproportionately escalating costs.

The rapid development of artificial intelligence and robotics has had a significant impact on our lives, with intelligent systems increasingly performing tasks traditionally performed by humans. Efficient knowledge transfer requires matching the mental model of the human teacher with the capabilities of the robot learner. This paper introduces the Mental Model Mismatch (MMM) Score, a feedback mechanism designed to quantify and reduce mismatches by aligning human teaching behavior with robot learning behavior. Using Large Language Models (LLMs), we analyze teacher intentions in natural language to generate adaptive feedback. A study with 150 participants teaching a virtual robot to solve a puzzle game shows that intention-based feedback significantly outperforms traditional performance-based feedback or no feedback. The results suggest that intention-based feedback improves instructional outcomes, improves understanding of the robot's learning process and reduces misconceptions. This research addresses a critical gap in human-robot interaction (HRI) by providing a method to quantify and mitigate discrepancies between human mental models and robot capabilities, with the goal of improving robot learning and human teaching effectiveness.

Abstract--Retrieval-augmented generation (RAG) enhances large language models (LLMs) by incorporating external knowledge to generate a response within a context with improved accuracy and reduced hallucinations. However, multi-modal RAG systems face unique challenges: (i) the retrieval process may select irrelevant entries to user query (e.g., images, documents), and (ii) vision-language models or multi-modal language models like GPT-4o may hallucinate when processing these entries to In this paper, we aim to address the first challenge, i.e, improving the selection of relevant context RAG. Specifically, we leverage the relevancy score (RS) measure designed in our previous work for evaluating the RAG performance The work in [4] large language models (LLMs) [2] by incorporating external employs multi-modal large language models (MLLMs) with knowledge to generate coherent responds based on a given knowledge-enhanced re-ranking and noise-injected training to context, improve the response accuracy and reduce hallucinations. Also, the work in [6] demonstrates the However, the quality of the generated response in RAG potential of vision-language models for relevance evaluation, systems heavily depends on the retrieval process. Selecting but it highlights these models' tendency to over-rely on the most relevant data from a database based on the user semantic similarity, often failing in cases that require precise query is essential for the system to generate accurate and contextual understanding. A common approach for approaches is their reduced ability to detect irrelevant retrieval in RAG is top-k selection by first ranking the entries information, as MLLMs are primarily trained on datasets from the knowledge-base based on similarity scores between containing relevant image-query pairs.

Effective instruction tuning is indispensable for optimizing code LLMs, aligning model behavior with user expectations and enhancing model performance in real-world applications. However, most existing methods focus on code snippets, which are limited to specific functionalities and rigid structures, restricting the complexity and diversity of the synthesized data. To address these limitations, we introduce a novel feature tree-based synthesis framework inspired by Abstract Syntax Trees (AST). Unlike AST, which captures syntactic structure of code, our framework models semantic relationships between code elements, enabling the generation of more nuanced and diverse data. The feature tree is constructed from raw data and refined iteratively to increase the quantity and diversity of the extracted features. This process enables the identification of more complex patterns and relationships within the code. By sampling subtrees with controlled depth and breadth, our framework allows precise adjustments to the complexity of the generated code, supporting a wide range of tasks from simple function-level operations to intricate multi-file scenarios. We fine-tuned widely-used base models to create the EpiCoder series, achieving state-of-the-art performance at both the function and file levels across multiple benchmarks. Notably, empirical evidence indicates that our approach shows significant potential in synthesizing highly complex repository-level code data. Further analysis elucidates the merits of this approach by rigorously assessing data complexity and diversity through software engineering principles and LLM-as-a-judge method.