Search

Large Language Models (LLMs) show potential for medical applications but often lack specialized clinical knowledge. Retrieval Augmented Generation (RAG) allows customization with domain-specific information, making it suitable for healthcare. This study evaluates the accuracy, consistency, and safety of RAG models in determining fitness for surgery and providing preoperative instructions. We developed LLM-RAG models using 35 local and 23 international preoperative guidelines and tested them against human-generated responses. A total of 3,682 responses were evaluated. Clinical documents were processed using Llamaindex, and 10 LLMs, including GPT3.5, GPT4, and Claude-3, were assessed. Fourteen clinical scenarios were analyzed, focusing on seven aspects of preoperative instructions. Established guidelines and expert judgment were used to determine correct responses, with human-generated answers serving as comparisons. The LLM-RAG models generated responses within 20 seconds, significantly faster than clinicians (10 minutes). The GPT4 LLM-RAG model achieved the highest accuracy (96.4% vs. 86.6%, p=0.016), with no hallucinations and producing correct instructions comparable to clinicians. Results were consistent across both local and international guidelines. This study demonstrates the potential of LLM-RAG models for preoperative healthcare tasks, highlighting their efficiency, scalability, and reliability., Comment: arXiv admin note: substantial text overlap with arXiv:2402.01733

This paper presents $\forall$uto$\exists$$\lor\!\land$L, a novel benchmark for scaling Large Language Model (LLM) assessment in formal tasks with clear notions of correctness, such as truth maintenance in translation and logical reasoning. $\forall$uto$\exists$$\lor\!\land$L is the first benchmarking paradigm that offers several key advantages necessary for scaling objective evaluation of LLMs without human labeling: (a) ability to evaluate LLMs of increasing sophistication by auto-generating tasks at different levels of difficulty; (b) auto-generation of ground truth that eliminates dependence on expensive and time-consuming human annotation; (c) the use of automatically generated, randomized datasets that mitigate the ability of successive LLMs to overfit to static datasets used in many contemporary benchmarks. Empirical analysis shows that an LLM's performance on $\forall$uto$\exists$$\lor\!\land$L is highly indicative of its performance on a diverse array of other benchmarks focusing on translation and reasoning tasks, making it a valuable autonomous evaluation paradigm in settings where hand-curated datasets can be hard to obtain and/or update.

When performing complex multi-step reasoning tasks, the ability of Large Language Models (LLMs) to derive structured intermediate proof steps is important for ensuring that the models truly perform the desired reasoning and for improving models' explainability. This paper is centred around a focused study: whether the current state-of-the-art generalist LLMs can leverage the structures in a few examples to better construct the proof structures with \textit{in-context learning}. Our study specifically focuses on structure-aware demonstration and structure-aware pruning. We demonstrate that they both help improve performance. A detailed analysis is provided to help understand the results., Comment: Accepted by EMNLP2024 main conference

The problem of symbolic music generation presents unique challenges due to the combination of limited data availability and the need for high precision in note pitch. To overcome these difficulties, we introduce Fine-grained Textural Guidance (FTG) within diffusion models to correct errors in the learned distributions. By incorporating FTG, the diffusion models improve the accuracy of music generation, which makes them well-suited for advanced tasks such as progressive music generation, improvisation and interactive music creation. We derive theoretical characterizations for both the challenges in symbolic music generation and the effect of the FTG approach. We provide numerical experiments and a demo page for interactive music generation with user input to showcase the effectiveness of our approach.

The emerging field of AI-generated art has witnessed the rise of prompt marketplaces, where creators can purchase, sell, or share prompts for generating unique artworks. These marketplaces often assert ownership over prompts, claiming them as intellectual property. This paper investigates whether concealed prompts sold on prompt marketplaces can be considered as secure intellectual property, given that humans and AI tools may be able to approximately infer the prompts based on publicly advertised sample images accompanying each prompt on sale. Specifically, our survey aims to assess (i) how accurately can humans infer the original prompt solely by examining an AI-generated image, with the goal of generating images similar to the original image, and (ii) the possibility of improving upon individual human and AI prompt inferences by crafting human-AI combined prompts with the help of a large language model. Although previous research has explored the use of AI and machine learning to infer (and also protect against) prompt inference, we are the first to include humans in the loop. Our findings indicate that while humans and human-AI collaborations can infer prompts and generate similar images with high accuracy, they are not as successful as using the original prompt.

Significant progress has been made in advancing large multimodal conversational models (LMMs), capitalizing on vast repositories of image-text data available online. Despite this progress, these models often encounter substantial domain gaps, hindering their ability to engage in complex conversations across new domains. Recent efforts have aimed to mitigate this issue, albeit relying on domain-specific image-text data to curate instruction-tuning data. However, many domains, such as agriculture, lack such vision-language data. In this work, we propose an approach to construct instruction-tuning data that harnesses vision-only data for the agriculture domain. We utilize diverse agricultural datasets spanning multiple domains, curate class-specific information, and employ large language models (LLMs) to construct an expert-tuning set, resulting in a 70k expert-tuning dataset called AgroInstruct. Subsequently, we expert-tuned and created AgroGPT, an efficient LMM that can hold complex agriculture-related conversations and provide useful insights. We also develop AgroEvals for evaluation and compare {AgroGPT's} performance with large open and closed-source models. {AgroGPT} excels at identifying fine-grained agricultural concepts, can act as an agriculture expert, and provides helpful information for multimodal agriculture questions. The code, datasets, and models are available at https://github.com/awaisrauf/agroGPT.

We present VoxelPrompt, an agent-driven vision-language framework that tackles diverse radiological tasks through joint modeling of natural language, image volumes, and analytical metrics. VoxelPrompt is multi-modal and versatile, leveraging the flexibility of language interaction while providing quantitatively grounded image analysis. Given a variable number of 3D medical volumes, such as MRI and CT scans, VoxelPrompt employs a language agent that iteratively predicts executable instructions to solve a task specified by an input prompt. These instructions communicate with a vision network to encode image features and generate volumetric outputs (e.g., segmentations). VoxelPrompt interprets the results of intermediate instructions and plans further actions to compute discrete measures (e.g., tumor growth across a series of scans) and present relevant outputs to the user. We evaluate this framework in a sandbox of diverse neuroimaging tasks, and we show that the single VoxelPrompt model can delineate hundreds of anatomical and pathological features, measure many complex morphological properties, and perform open-language analysis of lesion characteristics. VoxelPrompt carries out these objectives with accuracy similar to that of fine-tuned, single-task models for segmentation and visual question-answering, while facilitating a much larger range of tasks. Therefore, by supporting accurate image processing with language interaction, VoxelPrompt provides comprehensive utility for numerous imaging tasks that traditionally require specialized models to address., Comment: 21 pages, 5 figures, vision-language agent, medical image analysis, neuroimage foundation model

Relation extraction is crucial for constructing knowledge graphs, with large high-quality datasets serving as the foundation for training, fine-tuning, and evaluating models. Generative data augmentation (GDA) is a common approach to expand such datasets. However, this approach often introduces hallucinations, such as spurious facts, whose impact on relation extraction remains underexplored. In this paper, we examine the effects of hallucinations on the performance of relation extraction on the document and sentence levels. Our empirical study reveals that hallucinations considerably compromise the ability of models to extract relations from text, with recall reductions between 19.1% and 39.2%. We identify that relevant hallucinations impair the model's performance, while irrelevant hallucinations have a minimal impact. Additionally, we develop methods for the detection of hallucinations to improve data quality and model performance. Our approaches successfully classify texts as either 'hallucinated' or 'clean,' achieving high F1-scores of 83.8% and 92.2%. These methods not only assist in removing hallucinations but also help in estimating their prevalence within datasets, which is crucial for selecting high-quality data. Overall, our work confirms the profound impact of relevant hallucinations on the effectiveness of relation extraction models., Comment: Accepted at KBC-LM@ISWC'24

The detection of bias in natural language processing (NLP) is a critical challenge, particularly with the increasing use of large language models (LLMs) in various domains. This paper introduces GUS-Net, an innovative approach to bias detection that focuses on three key types of biases: (G)eneralizations, (U)nfairness, and (S)tereotypes. GUS-Net leverages generative AI and automated agents to create a comprehensive synthetic dataset, enabling robust multi-label token classification. Our methodology enhances traditional bias detection methods by incorporating the contextual encodings of pre-trained models, resulting in improved accuracy and depth in identifying biased entities. Through extensive experiments, we demonstrate that GUS-Net outperforms state-of-the-art techniques, achieving superior performance in terms of accuracy, F1-score, and Hamming Loss. The findings highlight GUS-Net's effectiveness in capturing a wide range of biases across diverse contexts, making it a valuable tool for social bias detection in text. This study contributes to the ongoing efforts in NLP to address implicit bias, providing a pathway for future research and applications in various fields. The Jupyter notebooks used to create the dataset and model are available at: https://github.com/Ethical-Spectacle/fair-ly/tree/main/resources. Warning: This paper contains examples of harmful language, and reader discretion is recommended.

Language models are extensively evaluated, but correctly interpreting evaluation results requires knowledge of train-test overlap which refers to the extent to which the language model is trained on the very data it is being tested on. The public currently lacks adequate information about train-test overlap: most models have no public train-test overlap statistics, and third parties cannot directly measure train-test overlap since they do not have access to the training data. To make this clear, we document the practices of 30 model developers, finding that just 9 developers report train-test overlap: 4 developers release training data under open-source licenses, enabling the community to directly measure train-test overlap, and 5 developers publish their train-test overlap methodology and statistics. By engaging with language model developers, we provide novel information about train-test overlap for three additional developers. Overall, we take the position that language model developers should publish train-test overlap statistics and/or training data whenever they report evaluation results on public test sets. We hope our work increases transparency into train-test overlap to increase the community-wide trust in model evaluations., Comment: 18 pages

Inference with transformer-based language models begins with a prompt processing step. In this step, the model generates the first output token and stores the KV cache needed for future generation steps. This prompt processing step can be computationally expensive, taking 10s of seconds or more for billion-parameter models on edge devices when prompt lengths or batch sizes rise. This degrades user experience by introducing significant latency into the model's outputs. To reduce the time spent producing the first output (known as the ``time to first token'', or TTFT) of a pretrained model, we introduce a novel method called KV Prediction. In our method, a small auxiliary model is used to process the prompt and produce an approximation of the KV cache used by a base model. This approximated KV cache is then used with the base model for autoregressive generation without the need to query the auxiliary model again. We demonstrate that our method produces a pareto-optimal efficiency-accuracy trade-off when compared to baselines. On TriviaQA, we demonstrate relative accuracy improvements in the range of $15\%-50\%$ across a range of TTFT FLOPs budgets. We also demonstrate accuracy improvements of up to $30\%$ on HumanEval python code completion at fixed TTFT FLOPs budgets. Additionally, we benchmark models on an Apple M2 Pro CPU and demonstrate that our improvement in FLOPs translates to a TTFT speedup on hardware. We release our code at https://github.com/apple/corenet/tree/main/projects/kv-prediction .

Graph-based anomaly detection is pivotal in diverse security applications, such as fraud detection in transaction networks and intrusion detection for network traffic. Standard approaches, including Graph Neural Networks (GNNs), often struggle to generalize across shifting data distributions. Meanwhile, real-world domain knowledge is more stable and a common existing component of real-world detection strategies. To explicitly integrate such knowledge into data-driven models such as GCNs, we propose KnowGraph, which integrates domain knowledge with data-driven learning for enhanced graph-based anomaly detection. KnowGraph comprises two principal components: (1) a statistical learning component that utilizes a main model for the overarching detection task, augmented by multiple specialized knowledge models that predict domain-specific semantic entities; (2) a reasoning component that employs probabilistic graphical models to execute logical inferences based on model outputs, encoding domain knowledge through weighted first-order logic formulas. Extensive experiments on these large-scale real-world datasets show that KnowGraph consistently outperforms state-of-the-art baselines in both transductive and inductive settings, achieving substantial gains in average precision when generalizing to completely unseen test graphs. Further ablation studies demonstrate the effectiveness of the proposed reasoning component in improving detection performance, especially under extreme class imbalance. These results highlight the potential of integrating domain knowledge into data-driven models for high-stakes, graph-based security applications., Comment: Accepted to ACM CCS 2024

Maternal mortality remains a significant global public health challenge. One promising approach to reducing maternal deaths occurring during facility-based childbirth is through early warning systems, which require the consistent monitoring of mothers' vital signs after giving birth. Wireless vital sign monitoring devices offer a labor-efficient solution for continuous monitoring, but their scarcity raises the critical question of how to allocate them most effectively. We devise an allocation algorithm for this problem by modeling it as a variant of the popular Restless Multi-Armed Bandit (RMAB) paradigm. In doing so, we identify and address novel, previously unstudied constraints unique to this domain, which render previous approaches for RMABs unsuitable and significantly increase the complexity of the learning and planning problem. To overcome these challenges, we adopt the popular Proximal Policy Optimization (PPO) algorithm from reinforcement learning to learn an allocation policy by training a policy and value function network. We demonstrate in simulations that our approach outperforms the best heuristic baseline by up to a factor of $4$.

By merging models, AI systems can combine the distinct strengths of separate language models, achieving a balance between multiple capabilities without requiring substantial retraining. However, the integration process can be intricate due to differences in training methods and fine-tuning, typically necessitating specialized knowledge and repeated refinement. This paper explores model merging techniques across a spectrum of complexity, examining where automated methods like evolutionary strategies stand compared to hyperparameter-driven approaches such as DARE, TIES-Merging and simpler methods like Model Soups. In addition, we introduce Differentiable Adaptive Merging (DAM), an efficient, adaptive merging approach as an alternative to evolutionary merging that optimizes model integration through scaling coefficients, minimizing computational demands. Our findings reveal that even simple averaging methods, like Model Soups, perform competitively when model similarity is high, underscoring each technique's unique strengths and limitations. We open-sourced DAM, including the implementation code and experiment pipeline, on GitHub: https://github.com/arcee-ai/DAM., Comment: 11 pages, 1 figure, and 3 tables

The improvement of economic policymaking presents an opportunity for broad societal benefit, a notion that has inspired research towards AI-driven policymaking tools. AI policymaking holds the potential to surpass human performance through the ability to process data quickly at scale. However, existing RL-based methods exhibit sample inefficiency, and are further limited by an inability to flexibly incorporate nuanced information into their decision-making processes. Thus, we propose a novel method in which we instead utilize pre-trained Large Language Models (LLMs), as sample-efficient policymakers in socially complex multi-agent reinforcement learning (MARL) scenarios. We demonstrate significant efficiency gains, outperforming existing methods across three environments. Our code is available at https://github.com/hegasz/large-legislative-models.

Images are capable of conveying emotions, but emotional experience is highly subjective. Advances in artificial intelligence have enabled the generation of images based on emotional descriptions. However, the level of agreement between the generative images and human emotional responses has not yet been evaluated. To address this, 20 artistic landscapes were generated using StyleGAN2-ADA. Four variants evoking positive emotions (contentment, amusement) and negative emotions (fear, sadness) were created for each image, resulting in 80 pictures. An online questionnaire was designed using this material, in which 61 observers classified the generated images. Statistical analyses were performed on the collected data to determine the level of agreement among participants, between the observer's responses, and the AI-generated emotions. A generally good level of agreement was found, with better results for negative emotions. However, the study confirms the subjectivity inherent in emotional evaluation., Comment: 29 pages

Banzhaf values offer a simple and interpretable alternative to the widely-used Shapley values. We introduce Kernel Banzhaf, a novel algorithm inspired by KernelSHAP, that leverages an elegant connection between Banzhaf values and linear regression. Through extensive experiments on feature attribution tasks, we demonstrate that Kernel Banzhaf substantially outperforms other algorithms for estimating Banzhaf values in both sample efficiency and robustness to noise. Furthermore, we prove theoretical guarantees on the algorithm's performance, establishing Kernel Banzhaf as a valuable tool for interpretable machine learning.

Large language models have enabled agents of all kinds to interact with users through natural conversation. Consequently, agents now have two jobs: conversing and planning/reasoning. Their conversational responses must be informed by all available information, and their actions must help to achieve goals. This dichotomy between conversing with the user and doing multi-step reasoning and planning can be seen as analogous to the human systems of "thinking fast and slow" as introduced by Kahneman. Our approach is comprised of a "Talker" agent (System 1) that is fast and intuitive, and tasked with synthesizing the conversational response; and a "Reasoner" agent (System 2) that is slower, more deliberative, and more logical, and is tasked with multi-step reasoning and planning, calling tools, performing actions in the world, and thereby producing the new agent state. We describe the new Talker-Reasoner architecture and discuss its advantages, including modularity and decreased latency. We ground the discussion in the context of a sleep coaching agent, in order to demonstrate real-world relevance.

This paper investigates how children learn numbers using the framework of reinforcement learning (RL), with a focus on the impact of language instructions. The motivation for using reinforcement learning stems from its parallels with psychological learning theories in controlled environments. By using state of the art deep reinforcement learning models, we simulate and analyze the effects of various forms of language instructions on number acquisition. Our findings indicate that certain linguistic structures more effectively improve numerical comprehension in RL agents. Additionally, our model predicts optimal sequences for presenting numbers to RL agents which enhance their speed of learning. This research provides valuable insights into the interplay between language and numerical cognition, with implications for both educational strategies and the development of artificial intelligence systems designed to support early childhood learning.

We address the problem of offline learning a policy that avoids undesirable demonstrations. Unlike conventional offline imitation learning approaches that aim to imitate expert or near-optimal demonstrations, our setting involves avoiding undesirable behavior (specified using undesirable demonstrations). To tackle this problem, unlike standard imitation learning where the aim is to minimize the distance between learning policy and expert demonstrations, we formulate the learning task as maximizing a statistical distance, in the space of state-action stationary distributions, between the learning policy and the undesirable policy. This significantly different approach results in a novel training objective that necessitates a new algorithm to address it. Our algorithm, UNIQ, tackles these challenges by building on the inverse Q-learning framework, framing the learning problem as a cooperative (non-adversarial) task. We then demonstrate how to efficiently leverage unlabeled data for practical training. Our method is evaluated on standard benchmark environments, where it consistently outperforms state-of-the-art baselines. The code implementation can be accessed at: https://github.com/hmhuy0/UNIQ.

The remarkable capability of Transformers to do reasoning and few-shot learning, without any fine-tuning, is widely conjectured to stem from their ability to implicitly simulate a multi-step algorithms -- such as gradient descent -- with their weights in a single forward pass. Recently, there has been progress in understanding this complex phenomenon from an expressivity point of view, by demonstrating that Transformers can express such multi-step algorithms. However, our knowledge about the more fundamental aspect of its learnability, beyond single layer models, is very limited. In particular, can training Transformers enable convergence to algorithmic solutions? In this work we resolve this for in-context linear regression with linear looped Transformers -- a multi-layer model with weight sharing that is conjectured to have an inductive bias to learn fix-point iterative algorithms. More specifically, for this setting we show that the global minimizer of the population training loss implements multi-step preconditioned gradient descent, with a preconditioner that adapts to the data distribution. Furthermore, we show a fast convergence for gradient flow on the regression loss, despite the non-convexity of the landscape, by proving a novel gradient dominance condition. To our knowledge, this is the first theoretical analysis for multi-layer Transformer in this setting. We further validate our theoretical findings through synthetic experiments.

Humans effortlessly integrate common-sense knowledge with sensory input from vision and touch to understand their surroundings. Emulating this capability, we introduce FusionSense, a novel 3D reconstruction framework that enables robots to fuse priors from foundation models with highly sparse observations from vision and tactile sensors. FusionSense addresses three key challenges: (i) How can robots efficiently acquire robust global shape information about the surrounding scene and objects? (ii) How can robots strategically select touch points on the object using geometric and common-sense priors? (iii) How can partial observations such as tactile signals improve the overall representation of the object? Our framework employs 3D Gaussian Splatting as a core representation and incorporates a hierarchical optimization strategy involving global structure construction, object visual hull pruning and local geometric constraints. This advancement results in fast and robust perception in environments with traditionally challenging objects that are transparent, reflective, or dark, enabling more downstream manipulation or navigation tasks. Experiments on real-world data suggest that our framework outperforms previously state-of-the-art sparse-view methods. All code and data are open-sourced on the project website.

As the cultural heritage sector increasingly adopts technologies like Retrieval-Augmented Generation (RAG) to provide more personalised search experiences and enable conversations with collections data, the demand for specialised evaluation datasets has grown. While end-to-end system testing is essential, it's equally important to assess individual components. We target the final, answering task, which is well-suited to Machine Reading Comprehension (MRC). Although existing MRC datasets address general domains, they lack the specificity needed for cultural heritage information. Unfortunately, the manual creation of such datasets is prohibitively expensive for most heritage institutions. This paper presents a cost-effective approach for generating domain-specific MRC datasets with increased difficulty using Reinforcement Learning from Human Feedback (RLHF) from synthetic preference data. Our method leverages the performance of existing question-answering models on a subset of SQuAD to create a difficulty metric, assuming that more challenging questions are answered correctly less frequently. This research contributes: (1) A methodology for increasing question difficulty using PPO and synthetic data; (2) Empirical evidence of the method's effectiveness, including human evaluation; (3) An in-depth error analysis and study of emergent phenomena; and (4) An open-source codebase and set of three llama-2-chat adapters for reproducibility and adaptation., Comment: is to be published in NLP4DH 2024

On-device adapting to continual, unpredictable domain shifts is essential for mobile applications like autonomous driving and augmented reality to deliver seamless user experiences in evolving environments. Test-time adaptation (TTA) emerges as a promising solution by tuning model parameters with unlabeled live data immediately before prediction. However, TTA's unique forward-backward-reforward pipeline notably increases the latency over standard inference, undermining the responsiveness in time-sensitive mobile applications. This paper presents AdaShadow, a responsive test-time adaptation framework for non-stationary mobile data distribution and resource dynamics via selective updates of adaptation-critical layers. Although the tactic is recognized in generic on-device training, TTA's unsupervised and online context presents unique challenges in estimating layer importance and latency, as well as scheduling the optimal layer update plan. AdaShadow addresses these challenges with a backpropagation-free assessor to rapidly identify critical layers, a unit-based runtime predictor to account for resource dynamics in latency estimation, and an online scheduler for prompt layer update planning. Also, AdaShadow incorporates a memory I/O-aware computation reuse scheme to further reduce latency in the reforward pass. Results show that AdaShadow achieves the best accuracy-latency balance under continual shifts. At low memory and energy costs, Adashadow provides a 2x to 3.5x speedup (ms-level) over state-of-the-art TTA methods with comparable accuracy and a 14.8% to 25.4% accuracy boost over efficient supervised methods with similar latency., Comment: This paper is accepted by SenSys 2024. Copyright may be transferred without notice

Motivated by interpretability and reliability, we investigate how neural networks represent knowledge during graph learning, We find hints of universality, where equivalent representations are learned across a range of model sizes (from $10^2$ to $10^9$ parameters) and contexts (MLP toy models, LLM in-context learning and LLM training). We show that these attractor representations optimize generalization to unseen examples by exploiting properties of knowledge graph relations (e.g. symmetry and meta-transitivity). We find experimental support for such universality by showing that LLMs and simpler neural networks can be stitched, i.e., by stitching the first part of one model to the last part of another, mediated only by an affine or almost affine transformation. We hypothesize that this dynamic toward simplicity and generalization is driven by "intelligence from starvation": where overfitting is minimized by pressure to minimize the use of resources that are either scarce or competed for against other tasks., Comment: 14 pages, 13 figures

As visual generation technologies continue to advance, the scale of video datasets has expanded rapidly, and the quality of these datasets is critical to the performance of video generation models. We argue that temporal splitting, detailed captions, and video quality filtering are three key factors that determine dataset quality. However, existing datasets exhibit various limitations in these areas. To address these challenges, we introduce Koala-36M, a large-scale, high-quality video dataset featuring accurate temporal splitting, detailed captions, and superior video quality. The core of our approach lies in improving the consistency between fine-grained conditions and video content. Specifically, we employ a linear classifier on probability distributions to enhance the accuracy of transition detection, ensuring better temporal consistency. We then provide structured captions for the splitted videos, with an average length of 200 words, to improve text-video alignment. Additionally, we develop a Video Training Suitability Score (VTSS) that integrates multiple sub-metrics, allowing us to filter high-quality videos from the original corpus. Finally, we incorporate several metrics into the training process of the generation model, further refining the fine-grained conditions. Our experiments demonstrate the effectiveness of our data processing pipeline and the quality of the proposed Koala-36M dataset. Our dataset and code will be released at https://koala36m.github.io/., Comment: Project page: https://koala36m.github.io/

With the rise of SSL and ASR technologies, the Wav2Vec2 ASR-based model has been fine-tuned for automated speech disorder quality assessment tasks, yielding impressive results and setting a new baseline for Head and Neck Cancer speech contexts. This demonstrates that the ASR dimension from Wav2Vec2 closely aligns with assessment dimensions. Despite its effectiveness, this system remains a black box with no clear interpretation of the connection between the model ASR dimension and clinical assessments. This paper presents the first analysis of this baseline model for speech quality assessment, focusing on intelligibility and severity tasks. We conduct a layer-wise analysis to identify key layers and compare different SSL and ASR Wav2Vec2 models based on pre-trained data. Additionally, post-hoc XAI methods, including Canonical Correlation Analysis (CCA) and visualization techniques, are used to track model evolution and visualize embeddings for enhanced interpretability., Comment: Accepted at the Spoken Language Technology (SLT) Conference 2024

Multimodal learning has gained increasing importance across various fields, offering the ability to integrate data from diverse sources such as images, text, and personalized records, which are frequently observed in medical domains. However, in scenarios where some modalities are missing, many existing frameworks struggle to accommodate arbitrary modality combinations, often relying heavily on a single modality or complete data. This oversight of potential modality combinations limits their applicability in real-world situations. To address this challenge, we propose Flex-MoE (Flexible Mixture-of-Experts), a new framework designed to flexibly incorporate arbitrary modality combinations while maintaining robustness to missing data. The core idea of Flex-MoE is to first address missing modalities using a new missing modality bank that integrates observed modality combinations with the corresponding missing ones. This is followed by a uniquely designed Sparse MoE framework. Specifically, Flex-MoE first trains experts using samples with all modalities to inject generalized knowledge through the generalized router ($\mathcal{G}$-Router). The $\mathcal{S}$-Router then specializes in handling fewer modality combinations by assigning the top-1 gate to the expert corresponding to the observed modality combination. We evaluate Flex-MoE on the ADNI dataset, which encompasses four modalities in the Alzheimer's Disease domain, as well as on the MIMIC-IV dataset. The results demonstrate the effectiveness of Flex-MoE highlighting its ability to model arbitrary modality combinations in diverse missing modality scenarios. Code is available at https://github.com/UNITES-Lab/flex-moe., Comment: NeurIPS 2024 Spotlight

Cross-domain recommendation (CDR) offers a promising solution to the data sparsity problem by enabling knowledge transfer across source and target domains. However, many recent CDR models overlook crucial issues such as privacy as well as the risk of negative transfer (which negatively impact model performance), especially in multi-domain settings. To address these challenges, we propose FedGCDR, a novel federated graph learning framework that securely and effectively leverages positive knowledge from multiple source domains. First, we design a positive knowledge transfer module that ensures privacy during inter-domain knowledge transmission. This module employs differential privacy-based knowledge extraction combined with a feature mapping mechanism, transforming source domain embeddings from federated graph attention networks into reliable domain knowledge. Second, we design a knowledge activation module to filter out potential harmful or conflicting knowledge from source domains, addressing the issues of negative transfer. This module enhances target domain training by expanding the graph of the target domain to generate reliable domain attentions and fine-tunes the target model for improved negative knowledge filtering and more accurate predictions. We conduct extensive experiments on 16 popular domains of the Amazon dataset, demonstrating that FedGCDR significantly outperforms state-of-the-art methods., Comment: Accepted by NeurIPS'24

Emergency department (ED) crowding is a global public health issue that has been repeatedly associated with increased mortality. Predicting future service demand would enable preventative measures aiming to eliminate crowding along with it's detrimental effects. Recent findings in our ED indicate that occupancy ratios exceeding 90% are associated with increased 10-day mortality. In this paper, we aim to predict these crisis periods using retrospective data from a large Nordic ED with a LightGBM model. We provide predictions for the whole ED and individually for it's different operational sections. We demonstrate that afternoon crowding can be predicted at 11 a.m. with an AUC of 0.82 (95% CI 0.78-0.86) and at 8 a.m. with an AUC up to 0.79 (95% CI 0.75-0.83). Consequently we show that forecasting mortality-associated crowding using anonymous administrative data is feasible.

Banking Transaction Flow (BTF) is a sequential data found in a number of banking activities such as marketing, credit risk or banking fraud. It is a multimodal data composed of three modalities: a date, a numerical value and a wording. We propose in this work an application of self-attention mechanism to the processing of BTFs. We trained two general models on a large amount of BTFs in a self-supervised way: one RNN-based model and one Transformer-based model. We proposed a specific tokenization in order to be able to process BTFs. The performance of these two models was evaluated on two banking downstream tasks: a transaction categorization task and a credit risk task. The results show that fine-tuning these two pre-trained models allowed to perform better than the state-of-the-art approaches for both tasks.

At the same time that artificial intelligence is becoming popular, concern and the need for regulation is growing, including among other requirements the data privacy. In this context, Federated Learning is proposed as a solution to data privacy concerns derived from different source data scenarios due to its distributed learning. The defense mechanisms proposed in literature are just focused on defending against adversarial attacks and the performance, leaving aside other important qualities such as explainability, fairness to poor quality clients, dynamism in terms of attacks configuration and generality in terms of being resilient against different kinds of attacks. In this work, we propose RAB$^2$-DEF, a $\textbf{r}$esilient $\textbf{a}$gainst $\textbf{b}\text{yzantine}$ and $\textbf{b}$ackdoor attacks which is $\textbf{d}$ynamic, $\textbf{e}$xplainable and $\textbf{f}$air to poor clients using local linear explanations. We test the performance of RAB$^2$-DEF in image datasets and both byzantine and backdoor attacks considering the state-of-the-art defenses and achieve that RAB$^2$-DEF is a proper defense at the same time that it boosts the other qualities towards trustworthy artificial intelligence.

Single point supervised oriented object detection has gained attention and made initial progress within the community. Diverse from those approaches relying on one-shot samples or powerful pretrained models (e.g. SAM), PointOBB has shown promise due to its prior-free feature. In this paper, we propose PointOBB-v2, a simpler, faster, and stronger method to generate pseudo rotated boxes from points without relying on any other prior. Specifically, we first generate a Class Probability Map (CPM) by training the network with non-uniform positive and negative sampling. We show that the CPM is able to learn the approximate object regions and their contours. Then, Principal Component Analysis (PCA) is applied to accurately estimate the orientation and the boundary of objects. By further incorporating a separation mechanism, we resolve the confusion caused by the overlapping on the CPM, enabling its operation in high-density scenarios. Extensive comparisons demonstrate that our method achieves a training speed 15.58x faster and an accuracy improvement of 11.60%/25.15%/21.19% on the DOTA-v1.0/v1.5/v2.0 datasets compared to the previous state-of-the-art, PointOBB. This significantly advances the cutting edge of single point supervised oriented detection in the modular track., Comment: 13 pages, 4 figures, 5 tables

Current large multimodal models (LMMs) face challenges in grounding, which requires the model to relate language components to visual entities. Contrary to the common practice that fine-tunes LMMs with additional grounding supervision, we find that the grounding ability can in fact emerge in LMMs trained without explicit grounding supervision. To reveal this emerging grounding, we introduce an "attend-and-segment" method which leverages attention maps from standard LMMs to perform pixel-level segmentation. Furthermore, to enhance the grounding ability, we propose DIFFLMM, an LMM utilizing a diffusion-based visual encoder, as opposed to the standard CLIP visual encoder, and trained with the same weak supervision. Without being constrained by the biases and limited scale of grounding-specific supervision data, our approach is more generalizable and scalable. We achieve competitive performance on both grounding-specific and general visual question answering benchmarks, compared with grounding LMMs and generalist LMMs, respectively. Notably, we achieve a 44.2 grounding mask recall on grounded conversation generation without any grounding supervision, outperforming the extensively supervised model GLaMM. Project page: https://groundLMM.github.io.

Large-scale vision-language pre-trained (VLP) models (e.g., CLIP) are renowned for their versatility, as they can be applied to diverse applications in a zero-shot setup. However, when these models are used in specific domains, their performance often falls short due to domain gaps or the under-representation of these domains in the training data. While fine-tuning VLP models on custom datasets with human-annotated labels can address this issue, annotating even a small-scale dataset (e.g., 100k samples) can be an expensive endeavor, often requiring expert annotators if the task is complex. To address these challenges, we propose LatteCLIP, an unsupervised method for fine-tuning CLIP models on classification with known class names in custom domains, without relying on human annotations. Our method leverages Large Multimodal Models (LMMs) to generate expressive textual descriptions for both individual images and groups of images. These provide additional contextual information to guide the fine-tuning process in the custom domains. Since LMM-generated descriptions are prone to hallucination or missing details, we introduce a novel strategy to distill only the useful information and stabilize the training. Specifically, we learn rich per-class prototype representations from noisy generated texts and dual pseudo-labels. Our experiments on 10 domain-specific datasets show that LatteCLIP outperforms pre-trained zero-shot methods by an average improvement of +4.74 points in top-1 accuracy and other state-of-the-art unsupervised methods by +3.45 points.

In this paper, we introduce SPA, a novel representation learning framework that emphasizes the importance of 3D spatial awareness in embodied AI. Our approach leverages differentiable neural rendering on multi-view images to endow a vanilla Vision Transformer (ViT) with intrinsic spatial understanding. We present the most comprehensive evaluation of embodied representation learning to date, covering 268 tasks across 8 simulators with diverse policies in both single-task and language-conditioned multi-task scenarios. The results are compelling: SPA consistently outperforms more than 10 state-of-the-art representation methods, including those specifically designed for embodied AI, vision-centric tasks, and multi-modal applications, while using less training data. Furthermore, we conduct a series of real-world experiments to confirm its effectiveness in practical scenarios. These results highlight the critical role of 3D spatial awareness for embodied representation learning. Our strongest model takes more than 6000 GPU hours to train and we are committed to open-sourcing all code and model weights to foster future research in embodied representation learning. Project Page: https://haoyizhu.github.io/spa/., Comment: Project Page: https://haoyizhu.github.io/spa/

Code has been shown to be effective in enhancing the mathematical reasoning abilities of large language models due to its precision and accuracy. Previous works involving continued mathematical pretraining often include code that utilizes math-related packages, which are primarily designed for fields such as engineering, machine learning, signal processing, or module testing, rather than being directly focused on mathematical reasoning. In this paper, we introduce a novel method for generating mathematical code accompanied with corresponding reasoning steps for continued pretraining. Our approach begins with the construction of a high-quality mathematical continued pretraining dataset by incorporating math-related web data, code using mathematical packages, math textbooks, and synthetic data. Next, we construct reasoning steps by extracting LaTeX expressions, the conditions needed for the expressions, and the results of the expressions from the previously collected dataset. Based on this extracted information, we generate corresponding code to accurately capture the mathematical reasoning process. Appending the generated code to each reasoning step results in data consisting of paired natural language reasoning steps and their corresponding code. Combining this data with the original dataset results in a 19.2B-token high-performing mathematical pretraining corpus, which we name MathCode-Pile. Training several popular base models with this corpus significantly improves their mathematical abilities, leading to the creation of the MathCoder2 family of models. All of our data processing and training code is open-sourced, ensuring full transparency and easy reproducibility of the entire data collection and training pipeline. The code is released at https://github.com/mathllm/MathCoder2 ., Comment: https://github.com/mathllm/MathCoder2

Tool learning enables Large Language Models (LLMs) to interact with external environments by invoking tools, serving as an effective strategy to mitigate the limitations inherent in their pre-training data. In this process, tool documentation plays a crucial role by providing usage instructions for LLMs, thereby facilitating effective tool utilization. This paper concentrates on the critical challenge of bridging the comprehension gap between LLMs and external tools due to the inadequacies and inaccuracies inherent in existing human-centric tool documentation. We propose a novel framework, DRAFT, aimed at Dynamically Refining tool documentation through the Analysis of Feedback and Trails emanating from LLMs' interactions with external tools. This methodology pivots on an innovative trial-and-error approach, consisting of three distinct learning phases: experience gathering, learning from experience, and documentation rewriting, to iteratively enhance the tool documentation. This process is further optimized by implementing a diversity-promoting exploration strategy to ensure explorative diversity and a tool-adaptive termination mechanism to prevent overfitting while enhancing efficiency. Extensive experiments on multiple datasets demonstrate that DRAFT's iterative, feedback-based refinement significantly ameliorates documentation quality, fostering a deeper comprehension and more effective utilization of tools by LLMs. Notably, our analysis reveals that the tool documentation refined via our approach demonstrates robust cross-model generalization capabilities.

We present a novel framework for free-viewpoint facial performance relighting using diffusion-based image-to-image translation. Leveraging a subject-specific dataset containing diverse facial expressions captured under various lighting conditions, including flat-lit and one-light-at-a-time (OLAT) scenarios, we train a diffusion model for precise lighting control, enabling high-fidelity relit facial images from flat-lit inputs. Our framework includes spatially-aligned conditioning of flat-lit captures and random noise, along with integrated lighting information for global control, utilizing prior knowledge from the pre-trained Stable Diffusion model. This model is then applied to dynamic facial performances captured in a consistent flat-lit environment and reconstructed for novel-view synthesis using a scalable dynamic 3D Gaussian Splatting method to maintain quality and consistency in the relit results. In addition, we introduce unified lighting control by integrating a novel area lighting representation with directional lighting, allowing for joint adjustments in light size and direction. We also enable high dynamic range imaging (HDRI) composition using multiple directional lights to produce dynamic sequences under complex lighting conditions. Our evaluations demonstrate the models efficiency in achieving precise lighting control and generalizing across various facial expressions while preserving detailed features such as skintexture andhair. The model accurately reproduces complex lighting effects like eye reflections, subsurface scattering, self-shadowing, and translucency, advancing photorealism within our framework., Comment: 18 pages, SIGGRAPH Asia 2024 Conference Papers (SA Conference Papers '24), December 3--6, 2024, Tokyo, Japan. Project page: https://www.eyelinestudios.com/research/diffrelight.html

Existing multimodal retrieval benchmarks primarily focus on evaluating whether models can retrieve and utilize external textual knowledge for question answering. However, there are scenarios where retrieving visual information is either more beneficial or easier to access than textual data. In this paper, we introduce a multimodal retrieval-augmented generation benchmark, MRAG-Bench, in which we systematically identify and categorize scenarios where visually augmented knowledge is better than textual knowledge, for instance, more images from varying viewpoints. MRAG-Bench consists of 16,130 images and 1,353 human-annotated multiple-choice questions across 9 distinct scenarios. With MRAG-Bench, we conduct an evaluation of 10 open-source and 4 proprietary large vision-language models (LVLMs). Our results show that all LVLMs exhibit greater improvements when augmented with images compared to textual knowledge, confirming that MRAG-Bench is vision-centric. Additionally, we conduct extensive analysis with MRAG-Bench, which offers valuable insights into retrieval-augmented LVLMs. Notably, the top-performing model, GPT-4o, faces challenges in effectively leveraging retrieved knowledge, achieving only a 5.82% improvement with ground-truth information, in contrast to a 33.16% improvement observed in human participants. These findings highlight the importance of MRAG-Bench in encouraging the community to enhance LVLMs' ability to utilize retrieved visual knowledge more effectively., Comment: https://mragbench.github.io

Multimodal Large Language Models (MLLMs) exhibit promising advancements across various tasks, yet they still encounter significant trustworthiness issues. Prior studies apply Split Conformal Prediction (SCP) in language modeling to construct prediction sets with statistical guarantees. However, these methods typically rely on internal model logits or are restricted to multiple-choice settings, which hampers their generalizability and adaptability in dynamic, open-ended environments. In this paper, we introduce TRON, a two-step framework for risk control and assessment, applicable to any MLLM that supports sampling in both open-ended and closed-ended scenarios. TRON comprises two main components: (1) a novel conformal score to sample response sets of minimum size, and (2) a nonconformity score to identify high-quality responses based on self-consistency theory, controlling the error rates by two specific risk levels. Furthermore, we investigate semantic redundancy in prediction sets within open-ended contexts for the first time, leading to a promising evaluation metric for MLLMs based on average set size. Our comprehensive experiments across four Video Question-Answering (VideoQA) datasets utilizing eight MLLMs show that TRON achieves desired error rates bounded by two user-specified risk levels. Additionally, deduplicated prediction sets maintain adaptiveness while being more efficient and stable for risk assessment under different risk levels., Comment: 15 pages, 6 figures

Due to the difficulty of acquiring extensive real-world data, robot simulation has become crucial for parallel training and sim-to-real transfer, highlighting the importance of scalable simulated robotic tasks. Foundation models have demonstrated impressive capacities in autonomously generating feasible robotic tasks. However, this new paradigm underscores the challenge of adequately evaluating these autonomously generated tasks. To address this, we propose a comprehensive evaluation framework tailored to generative simulations. Our framework segments evaluation into three core aspects: quality, diversity, and generalization. For single-task quality, we evaluate the realism of the generated task and the completeness of the generated trajectories using large language models and vision-language models. In terms of diversity, we measure both task and data diversity through text similarity of task descriptions and world model loss trained on collected task trajectories. For task-level generalization, we assess the zero-shot generalization ability on unseen tasks of a policy trained with multiple generated tasks. Experiments conducted on three representative task generation pipelines demonstrate that the results from our framework are highly consistent with human evaluations, confirming the feasibility and validity of our approach. The findings reveal that while metrics of quality and diversity can be achieved through certain methods, no single approach excels across all metrics, suggesting a need for greater focus on balancing these different metrics. Additionally, our analysis further highlights the common challenge of low generalization capability faced by current works. Our anonymous website: https://sites.google.com/view/evaltasks., Comment: Project website: https://sites.google.com/view/evaltasks

We present Agent S, an open agentic framework that enables autonomous interaction with computers through a Graphical User Interface (GUI), aimed at transforming human-computer interaction by automating complex, multi-step tasks. Agent S aims to address three key challenges in automating computer tasks: acquiring domain-specific knowledge, planning over long task horizons, and handling dynamic, non-uniform interfaces. To this end, Agent S introduces experience-augmented hierarchical planning, which learns from external knowledge search and internal experience retrieval at multiple levels, facilitating efficient task planning and subtask execution. In addition, it employs an Agent-Computer Interface (ACI) to better elicit the reasoning and control capabilities of GUI agents based on Multimodal Large Language Models (MLLMs). Evaluation on the OSWorld benchmark shows that Agent S outperforms the baseline by 9.37% on success rate (an 83.6% relative improvement) and achieves a new state-of-the-art. Comprehensive analysis highlights the effectiveness of individual components and provides insights for future improvements. Furthermore, Agent S demonstrates broad generalizability to different operating systems on a newly-released WindowsAgentArena benchmark. Code available at https://github.com/simular-ai/Agent-S., Comment: 23 pages, 16 figures, 9 tables

Large language models (LLMs) have achieved reasonable quality improvements in machine translation (MT). However, most current research on MT-LLMs still faces significant challenges in maintaining translation consistency and accuracy when processing entire documents. In this paper, we introduce DelTA, a Document-levEL Translation Agent designed to overcome these limitations. DelTA features a multi-level memory structure that stores information across various granularities and spans, including Proper Noun Records, Bilingual Summary, Long-Term Memory, and Short-Term Memory, which are continuously retrieved and updated by auxiliary LLM-based components. Experimental results indicate that DelTA significantly outperforms strong baselines in terms of translation consistency and quality across four open/closed-source LLMs and two representative document translation datasets, achieving an increase in consistency scores by up to 4.58 percentage points and in COMET scores by up to 3.16 points on average. DelTA employs a sentence-by-sentence translation strategy, ensuring no sentence omissions and offering a memory-efficient solution compared to the mainstream method. Furthermore, DelTA improves pronoun translation accuracy, and the summary component of the agent also shows promise as a tool for query-based summarization tasks. We release our code and data at https://github.com/YutongWang1216/DocMTAgent.

Generative modeling of discrete data underlies important applications spanning text-based agents like ChatGPT to the design of the very building blocks of life in protein sequences. However, application domains need to exert control over the generated data by steering the generative process - typically via RLHF - to satisfy a specified property, reward, or affinity metric. In this paper, we study the problem of steering Masked Diffusion Models (MDMs), a recent class of discrete diffusion models that offer a compelling alternative to traditional autoregressive models. We introduce Discrete Denoising Posterior Prediction (DDPP), a novel framework that casts the task of steering pre-trained MDMs as a problem of probabilistic inference by learning to sample from a target Bayesian posterior. Our DDPP framework leads to a family of three novel objectives that are all simulation-free, and thus scalable while applying to general non-differentiable reward functions. Empirically, we instantiate DDPP by steering MDMs to perform class-conditional pixel-level image modeling, RLHF-based alignment of MDMs using text-based rewards, and finetuning protein language models to generate more diverse secondary structures and shorter proteins. We substantiate our designs via wet-lab validation, where we observe transient expression of reward-optimized protein sequences.

Large Language Models (LLMs) trained on massive corpora have shown remarkable success in knowledge-intensive tasks. Yet, most of them rely on pre-stored knowledge. Inducing new general knowledge from a specific environment and performing reasoning with the acquired knowledge -- \textit{situated inductive reasoning}, is crucial and challenging for machine intelligence. In this paper, we design Mars, an interactive environment devised for situated inductive reasoning. It introduces counter-commonsense game mechanisms by modifying terrain, survival setting and task dependency while adhering to certain principles. In Mars, agents need to actively interact with their surroundings, derive useful rules and perform decision-making tasks in specific contexts. We conduct experiments on various RL-based and LLM-based methods, finding that they all struggle on this challenging situated inductive reasoning benchmark. Furthermore, we explore \textit{Induction from Reflection}, where we instruct agents to perform inductive reasoning from history trajectory. The superior performance underscores the importance of inductive reasoning in Mars. Through Mars, we aim to galvanize advancements in situated inductive reasoning and set the stage for developing the next generation of AI systems that can reason in an adaptive and context-sensitive way.

Current LLM benchmarks focus on evaluating models' memory of facts and semantic relations, primarily assessing semantic aspects of long-term memory. However, in humans, long-term memory also includes episodic memory, which links memories to their contexts, such as the time and place they occurred. The ability to contextualize memories is crucial for many cognitive tasks and everyday functions. This form of memory has not been evaluated in LLMs with existing benchmarks. To address the gap in evaluating memory in LLMs, we introduce Sequence Order Recall Tasks (SORT), which we adapt from tasks used to study episodic memory in cognitive psychology. SORT requires LLMs to recall the correct order of text segments, and provides a general framework that is both easily extendable and does not require any additional annotations. We present an initial evaluation dataset, Book-SORT, comprising 36k pairs of segments extracted from 9 books recently added to the public domain. Based on a human experiment with 155 participants, we show that humans can recall sequence order based on long-term memory of a book. We find that models can perform the task with high accuracy when relevant text is given in-context during the SORT evaluation. However, when presented with the book text only during training, LLMs' performance on SORT falls short. By allowing to evaluate more aspects of memory, we believe that SORT will aid in the emerging development of memory-augmented models.

Residential rooftop solar adoption is considered crucial for reducing carbon emissions. The lack of photovoltaic (PV) data at a finer resolution (e.g., household, hourly levels) poses a significant roadblock to informed decision-making. We discuss a novel methodology to generate a highly granular, residential-scale realistic dataset for rooftop solar adoption across the contiguous United States. The data-driven methodology consists of: (i) integrated machine learning models to identify PV adopters, (ii) methods to augment the data using explainable AI techniques to glean insights about key features and their interactions, and (iii) methods to generate household-level hourly solar energy output using an analytical model. The resulting synthetic datasets are validated using real-world data and can serve as a digital twin for modeling downstream tasks. Finally, a policy-based case study utilizing the digital twin for Virginia demonstrated increased rooftop solar adoption with the 30\% Federal Solar Investment Tax Credit, especially in Low-to-Moderate-Income communities., Comment: 41 pages including references and supplementary

The digital revolution has significantly impacted financial transactions, leading to a notable increase in credit card usage. However, this convenience comes with a trade-off: a substantial rise in fraudulent activities. Traditional machine learning methods for fraud detection often struggle to capture the inherent interconnectedness within financial data. This paper proposes a novel approach for credit card fraud detection that leverages Graph Neural Networks (GNNs) with attention mechanisms applied to heterogeneous graph representations of financial data. Unlike homogeneous graphs, heterogeneous graphs capture intricate relationships between various entities in the financial ecosystem, such as cardholders, merchants, and transactions, providing a richer and more comprehensive data representation for fraud analysis. To address the inherent class imbalance in fraud data, where genuine transactions significantly outnumber fraudulent ones, the proposed approach integrates an autoencoder. This autoencoder, trained on genuine transactions, learns a latent representation and flags deviations during reconstruction as potential fraud. This research investigates two key questions: (1) How effectively can a GNN with an attention mechanism detect and prevent credit card fraud when applied to a heterogeneous graph? (2) How does the efficacy of the autoencoder with attention approach compare to traditional methods? The results are promising, demonstrating that the proposed model outperforms benchmark algorithms such as Graph Sage and FI-GRL, achieving a superior AUC-PR of 0.89 and an F1-score of 0.81. This research significantly advances fraud detection systems and the overall security of financial transactions by leveraging GNNs with attention mechanisms and addressing class imbalance through an autoencoder.

Large Language Model (LLM) based multi-agent systems (MAS) show remarkable potential in collaborative problem-solving, yet they still face critical challenges: low communication efficiency, poor scalability, and a lack of effective parameter-updating optimization methods. We present Optima, a novel framework that addresses these issues by significantly enhancing both communication efficiency and task effectiveness in LLM-based MAS through LLM training. Optima employs an iterative generate, rank, select, and train paradigm with a reward function balancing task performance, token efficiency, and communication readability. We explore various RL algorithms, including Supervised Fine-Tuning, Direct Preference Optimization, and their hybrid approaches, providing insights into their effectiveness-efficiency trade-offs. We integrate Monte Carlo Tree Search-inspired techniques for DPO data generation, treating conversation turns as tree nodes to explore diverse interaction paths. Evaluated on common multi-agent tasks, including information-asymmetric question answering and complex reasoning, Optima shows consistent and substantial improvements over single-agent baselines and vanilla MAS based on Llama 3 8B, achieving up to 2.8x performance gain with less than 10\% tokens on tasks requiring heavy information exchange. Moreover, Optima's efficiency gains open new possibilities for leveraging inference-compute more effectively, leading to improved inference-time scaling laws. By addressing fundamental challenges in LLM-based MAS, Optima shows the potential towards scalable, efficient, and effective MAS (https://chenweize1998.github.io/optima-project-page)., Comment: Under review