Jordan W. Suchow

Jimin Huang, Mengxi Xiao, Dong Li et al.

Financial LLMs hold promise for advancing financial tasks and domain-specific applications. However, they are limited by scarce corpora, weak multimodal capabilities, and narrow evaluations, making them less suited for real-world application. To address this, we introduce \textit{Open-FinLLMs}, the first open-source multimodal financial LLMs designed to handle diverse tasks across text, tabular, time-series, and chart data, excelling in zero-shot, few-shot, and fine-tuning settings. The suite includes FinLLaMA, pre-trained on a comprehensive 52-billion-token corpus; FinLLaMA-Instruct, fine-tuned with 573K financial instructions; and FinLLaVA, enhanced with 1.43M multimodal tuning pairs for strong cross-modal reasoning. We comprehensively evaluate Open-FinLLMs across 14 financial tasks, 30 datasets, and 4 multimodal tasks in zero-shot, few-shot, and supervised fine-tuning settings, introducing two new multimodal evaluation datasets. Our results show that Open-FinLLMs outperforms afvanced financial and general LLMs such as GPT-4, across financial NLP, decision-making, and multi-modal tasks, highlighting their potential to tackle real-world challenges. To foster innovation and collaboration across academia and industry, we release all codes (https://anonymous.4open.science/r/PIXIU2-0D70/B1D7/LICENSE) and models under OSI-approved licenses.

Yangyang Yu, Haohang Li, Zhi Chen et al.

Recent advancements in Large Language Models (LLMs) have exhibited notable efficacy in question-answering (QA) tasks across diverse domains. Their prowess in integrating extensive web knowledge has fueled interest in developing LLM-based autonomous agents. While LLMs are efficient in decoding human instructions and deriving solutions by holistically processing historical inputs, transitioning to purpose-driven agents requires a supplementary rational architecture to process multi-source information, establish reasoning chains, and prioritize critical tasks. Addressing this, we introduce \textsc{FinMem}, a novel LLM-based agent framework devised for financial decision-making. It encompasses three core modules: Profiling, to customize the agent's characteristics; Memory, with layered message processing, to aid the agent in assimilating hierarchical financial data; and Decision-making, to convert insights gained from memories into investment decisions. Notably, \textsc{FinMem}'s memory module aligns closely with the cognitive structure of human traders, offering robust interpretability and real-time tuning. Its adjustable cognitive span allows for the retention of critical information beyond human perceptual limits, thereby enhancing trading outcomes. This framework enables the agent to self-evolve its professional knowledge, react agilely to new investment cues, and continuously refine trading decisions in the volatile financial environment. We first compare \textsc{FinMem} with various algorithmic agents on a scalable real-world financial dataset, underscoring its leading trading performance in stocks. We then fine-tuned the agent's perceptual span and character setting to achieve a significantly enhanced trading performance. Collectively, \textsc{FinMem} presents a cutting-edge LLM agent framework for automated trading, boosting cumulative investment returns.

Yangyang Yu, Zhiyuan Yao, Haohang Li et al.

Large language models (LLMs) have demonstrated notable potential in conducting complex tasks and are increasingly utilized in various financial applications. However, high-quality sequential financial investment decision-making remains challenging. These tasks require multiple interactions with a volatile environment for every decision, demanding sufficient intelligence to maximize returns and manage risks. Although LLMs have been used to develop agent systems that surpass human teams and yield impressive investment returns, opportunities to enhance multi-sourced information synthesis and optimize decision-making outcomes through timely experience refinement remain unexplored. Here, we introduce the FinCon, an LLM-based multi-agent framework with CONceptual verbal reinforcement tailored for diverse FINancial tasks. Inspired by effective real-world investment firm organizational structures, FinCon utilizes a manager-analyst communication hierarchy. This structure allows for synchronized cross-functional agent collaboration towards unified goals through natural language interactions and equips each agent with greater memory capacity than humans. Additionally, a risk-control component in FinCon enhances decision quality by episodically initiating a self-critiquing mechanism to update systematic investment beliefs. The conceptualized beliefs serve as verbal reinforcement for the future agent's behavior and can be selectively propagated to the appropriate node that requires knowledge updates. This feature significantly improves performance while reducing unnecessary peer-to-peer communication costs. Moreover, FinCon demonstrates strong generalization capabilities in various financial tasks, including single stock trading and portfolio management.

Yangyang Yu, Jordan W. Suchow

High-dimensional deep neural network representations of images and concepts can be aligned to predict human annotations of diverse stimuli. However, such alignment requires the costly collection of behavioral responses, such that, in practice, the deep-feature spaces are only ever sparsely sampled. Here, we propose an active learning approach to adaptively sampling experimental stimuli to efficiently learn a Bayesian matrix factorization model with deep side information. We observe a significant efficiency gain over a passive baseline. Furthermore, with a sequential batched sampling strategy, the algorithm is applicable not only to small datasets collected from traditional laboratory experiments but also to settings where large-scale crowdsourced data collection is needed to accurately align the high-dimensional deep feature representations derived from pre-trained networks.

Yupeng Cao, Haohang Li, Weijin Liu et al.

Recent studies demonstrate that tool-calling capability enables large language models (LLMs) to interact with external environments for long-horizon financial tasks. While existing benchmarks have begun evaluating financial tool calling, they focus on limited scenarios and rely on call-level metrics that fail to capture trajectory-level reasoning quality. To address this gap, we introduce FinTrace, a benchmark comprising 800 expert-annotated trajectories spanning 34 real-world financial task categories across multiple difficulty levels. FinTrace employs a rubric-based evaluation protocol with nine metrics organized along four axes -- action correctness, execution efficiency, process quality, and output quality -- enabling fine-grained assessment of LLM tool-calling behavior. Our evaluation of 13 LLMs reveals that while frontier models achieve strong tool selection, all models struggle with information utilization and final answer quality, exposing a critical gap between invoking the right tools and reasoning effectively over their outputs. To move beyond diagnosis, we construct FinTrace-Training, the first trajectory-level preference dataset for financial tool-calling, containing 8,196 curated trajectories with tool-augmented contexts and preference pairs. We fine-tune Qwen-3.5-9B using supervised fine-tuning followed by direct preference optimization (DPO) and show that training on FinTrace-Training consistently improves intermediate reasoning metrics, with DPO more effectively suppressing failure modes. However, end-to-end answer quality remains a bottleneck, indicating that trajectory-level improvements do not yet fully propagate to final output quality.

Necdet Gürkan, Jordan W. Suchow

Harnessing collective intelligence to drive effective decision-making and collaboration benefits from the ability to detect and characterize heterogeneity in consensus beliefs. This is particularly true in domains such as technology acceptance or leadership perception, where a consensus defines an intersubjective truth, leading to the possibility of multiple "ground truths" when subsets of respondents sustain mutually incompatible consensuses. Cultural Consensus Theory (CCT) provides a statistical framework for detecting and characterizing these divergent consensus beliefs. However, it is unworkable in modern applications because it lacks the ability to generalize across even highly similar beliefs, is ineffective with sparse data, and can leverage neither external knowledge bases nor learned machine representations. Here, we overcome these limitations through Infinite Deep Latent Construct Cultural Consensus Theory (iDLC-CCT), a nonparametric Bayesian model that extends CCT with a latent construct that maps between pretrained deep neural network embeddings of entities and the consensus beliefs regarding those entities among one or more subsets of respondents. We validate the method across domains including perceptions of risk sources, food healthiness, leadership, first impressions, and humor. We find that iDLC-CCT better predicts the degree of consensus, generalizes well to out-of-sample entities, and is effective even with sparse data. To improve scalability, we introduce an efficient hard-clustering variant of the iDLC-CCT using an algorithm derived from a small-variance asymptotic analysis of the model. The iDLC-CCT, therefore, provides a workable computational foundation for harnessing collective intelligence under a lack of cultural consensus and may potentially form the basis of consensus-aware information technologies.

Haohang Li, Yupeng Cao, Yangyang Yu et al. · utoronto

Recent advancements have underscored the potential of large language model (LLM)-based agents in financial decision-making. Despite this progress, the field currently encounters two main challenges: (1) the lack of a comprehensive LLM agent framework adaptable to a variety of financial tasks, and (2) the absence of standardized benchmarks and consistent datasets for assessing agent performance. To tackle these issues, we introduce \textsc{InvestorBench}, the first benchmark specifically designed for evaluating LLM-based agents in diverse financial decision-making contexts. InvestorBench enhances the versatility of LLM-enabled agents by providing a comprehensive suite of tasks applicable to different financial products, including single equities like stocks, cryptocurrencies and exchange-traded funds (ETFs). Additionally, we assess the reasoning and decision-making capabilities of our agent framework using thirteen different LLMs as backbone models, across various market environments and tasks. Furthermore, we have curated a diverse collection of open-source, multi-modal datasets and developed a comprehensive suite of environments for financial decision-making. This establishes a highly accessible platform for evaluating financial agents' performance across various scenarios.

Jordan W. Suchow, Necdet Gürkan

Generative A.I. models have emerged as versatile tools across diverse industries, with applications in privacy-preserving data sharing, computational art, personalization of products and services, and immersive entertainment. Here, we introduce a new privacy concern in the adoption and use of generative A.I. models: that of coincidental generation, where a generative model's output is similar enough to an existing entity, beyond those represented in the dataset used to train the model, to be mistaken for it. Consider, for example, synthetic portrait generators, which are today deployed in commercial applications such as virtual modeling agencies and synthetic stock photography. Due to the low intrinsic dimensionality of human face perception, every synthetically generated face will coincidentally resemble an actual person. Such examples of coincidental generation all but guarantee the misappropriation of likeness and expose organizations that use generative A.I. to legal and regulatory risk.

Vahid Ashrafimoghari, Necdet Gürkan, Jordan W. Suchow

The rapid evolution of artificial intelligence (AI), especially in the domain of Large Language Models (LLMs) and generative AI, has opened new avenues for application across various fields, yet its role in business education remains underexplored. This study introduces the first benchmark to assess the performance of seven major LLMs, OpenAI's models (GPT-3.5 Turbo, GPT-4, and GPT-4 Turbo), Google's models (PaLM 2, Gemini 1.0 Pro), and Anthropic's models (Claude 2 and Claude 2.1), on the GMAT, which is a key exam in the admission process for graduate business programs. Our analysis shows that most LLMs outperform human candidates, with GPT-4 Turbo not only outperforming the other models but also surpassing the average scores of graduate students at top business schools. Through a case study, this research examines GPT-4 Turbo's ability to explain answers, evaluate responses, identify errors, tailor instructions, and generate alternative scenarios. The latest LLM versions, GPT-4 Turbo, Claude 2.1, and Gemini 1.0 Pro, show marked improvements in reasoning tasks compared to their predecessors, underscoring their potential for complex problem-solving. While AI's promise in education, assessment, and tutoring is clear, challenges remain. Our study not only sheds light on LLMs' academic potential but also emphasizes the need for careful development and application of AI in education. As AI technology advances, it is imperative to establish frameworks and protocols for AI interaction, verify the accuracy of AI-generated content, ensure worldwide access for diverse learners, and create an educational environment where AI supports human expertise. This research sets the stage for further exploration into the responsible use of AI to enrich educational experiences and improve exam preparation and assessment methods.

Necdet Gurkan, Jordan W. Suchow

As the societal implications of Artificial Intelligence (AI) continue to grow, the pursuit of responsible AI necessitates public engagement in its development and governance processes. This involvement is crucial for capturing diverse perspectives and promoting equitable practices and outcomes. We applied Cultural Consensus Theory (CCT) to a nationally representative survey dataset on various aspects of AI to discern beliefs and attitudes about responsible AI in the United States. Our results offer valuable insights by identifying shared and contrasting views on responsible AI. Furthermore, these findings serve as critical reference points for developers and policymakers, enabling them to more effectively consider individual variances and group-level cultural perspectives when making significant decisions and addressing the public's concerns.

Haohang Li, Yupeng Cao, Yangyang Yu et al.

Despite their remarkable success and deployment across diverse workflows, language models sometimes produce untruthful responses. Our limited understanding of how truthfulness is mechanistically encoded within these models jeopardizes their reliability and safety. In this paper, we propose a method for identifying representations of truthfulness at the neuron level. We show that language models contain truth neurons, which encode truthfulness in a subject-agnostic manner. Experiments conducted across models of varying scales validate the existence of truth neurons, confirming that the encoding of truthfulness at the neuron level is a property shared by many language models. The distribution patterns of truth neurons over layers align with prior findings on the geometry of truthfulness. Selectively suppressing the activations of truth neurons found through the TruthfulQA dataset degrades performance both on TruthfulQA and on other benchmarks, showing that the truthfulness mechanisms are not tied to a specific dataset. Our results offer novel insights into the mechanisms underlying truthfulness in language models and highlight potential directions toward improving their trustworthiness and reliability.

Necdet Gurkan, Kimathi Njoki, Jordan W. Suchow

As artificial intelligence (AI) continues to advance--particularly in generative models--an open question is whether these systems can replicate foundational models of human social perception. A well-established framework in social cognition suggests that social judgments are organized along two primary dimensions: valence (e.g., trustworthiness, warmth) and dominance (e.g., power, assertiveness). This study examines whether multimodal generative AI systems can reproduce this valence-dominance structure when evaluating facial images and how their representations align with those observed across world regions. Through principal component analysis (PCA), we found that the extracted dimensions closely mirrored the theoretical structure of valence and dominance, with trait loadings aligning with established definitions. However, many world regions and generative AI models also exhibited a third component, the nature and significance of which warrant further investigation. These findings demonstrate that multimodal generative AI systems can replicate key aspects of human social perception, raising important questions about their implications for AI-driven decision-making and human-AI interactions.

Necdet Gurkan, Jordan W. Suchow

Face detection is a long-standing challenge in the field of computer vision, with the ultimate goal being to accurately localize human faces in an unconstrained environment. There are significant technical hurdles in making these systems accurate due to confounding factors related to pose, image resolution, illumination, occlusion, and viewpoint [44]. That being said, with recent developments in machine learning, face-detection systems have achieved extraordinary accuracy, largely built on data-driven deep-learning models [70]. Though encouraging, a critical aspect that limits face-detection performance and social responsibility of deployed systems is the inherent diversity of human appearance. Every human appearance reflects something unique about a person, including their heritage, identity, experiences, and visible manifestations of self-expression. However, there are questions about how well face-detection systems perform when faced with varying face size and shape, skin color, body modification, and body ornamentation. Towards this goal, we collected the Distinctive Human Appearance dataset, an image set that represents appearances with low frequency and that tend to be undersampled in face datasets. Then, we evaluated current state-of-the-art face-detection models in their ability to detect faces in these images. The evaluation results show that face-detection algorithms do not generalize well to these diverse appearances. Evaluating and characterizing the state of current face-detection models will accelerate research and development towards creating fairer and more accurate face-detection systems.

Max Simchowitz, Kevin Jamieson, Jordan W. Suchow et al.

In this paper, we introduce the first principled adaptive-sampling procedure for learning a convex function in the $L_\infty$ norm, a problem that arises often in the behavioral and social sciences. We present a function-specific measure of complexity and use it to prove that, for each convex function $f_{\star}$, our algorithm nearly attains the information-theoretically optimal, function-specific error rate. We also corroborate our theoretical contributions with numerical experiments, finding that our method substantially outperforms passive, uniform sampling for favorable synthetic and data-derived functions in low-noise settings with large sampling budgets. Our results also suggest an idealized "oracle strategy", which we use to gauge the potential advance of any adaptive-sampling strategy over passive sampling, for any given convex function.

Jordan W. Suchow, Joshua C. Peterson, Thomas L. Griffiths

Generative models of human identity and appearance have broad applicability to behavioral science and technology, but the exquisite sensitivity of human face perception means that their utility hinges on the alignment of the model's representation to human psychological representations and the photorealism of the generated images. Meeting these requirements is an exacting task, and existing models of human identity and appearance are often unworkably abstract, artificial, uncanny, or biased. Here, we use a variational autoencoder with an autoregressive decoder to learn a face space from a uniquely diverse dataset of portraits that control much of the variation irrelevant to human identity and appearance. Our method generates photorealistic portraits of fictive identities with a smooth, navigable latent space. We validate our model's alignment with human sensitivities by introducing a psychophysical Turing test for images, which humans mostly fail. Lastly, we demonstrate an initial application of our model to the problem of fast search in mental space to obtain detailed "police sketches" in a small number of trials.

Joshua C. Peterson, Jordan W. Suchow, Krisha Aghi et al.

Understanding how people represent categories is a core problem in cognitive science. Decades of research have yielded a variety of formal theories of categories, but validating them with naturalistic stimuli is difficult. The challenge is that human category representations cannot be directly observed and running informative experiments with naturalistic stimuli such as images requires a workable representation of these stimuli. Deep neural networks have recently been successful in solving a range of computer vision tasks and provide a way to compactly represent image features. Here, we introduce a method to estimate the structure of human categories that combines ideas from cognitive science and machine learning, blending human-based algorithms with state-of-the-art deep image generators. We provide qualitative and quantitative results as a proof-of-concept for the method's feasibility. Samples drawn from human distributions rival those from state-of-the-art generative models in quality and outperform alternative methods for estimating the structure of human categories.