arXiv subjects

Zhuo Cai

Publications and source records attributed to Zhuo Cai.

8 recordsLinked to original sources

arXiv2026

When and How Order-Flow Interventions Work After Dealers Adapt

Learning dealer agents can settle on persistently wide quotes, raising what customers pay to trade. A venue can route more orders to tighter quotes, but a wide quote may reflect real risk, and once dealers adapt, an end-to-end comparison cannot show why a rule works: because it intervenes more, picks better moments, is announced in advance, or rewards the right quote. We propose SOFA (State-aligned Order-Flow Allocation), a public order-flow contract with three parts: a State Scorer ranks market states from lagged public data, a Commitment Scheduler announces a budgeted intervention strength before dealers quote, and a Quote-to-Flow Allocator shifts a bounded share of fixed customer demand toward tighter quotes. Our paired re-adaptation protocol gives competing rules the same intervention strengths, only re-timed, lets dealers re-adapt to each rule from a shared checkpoint, and breaks one link at a time. Across 20 independently trained dealer populations in an ABIDES market, concentrating the intervention in states that the score ranks as low-risk lowers low-risk execution cost relative to random timing in every population, by 0.040 tick on average, while a stale announcement or a broken quote-to-flow link raises this cost. In the primary setting the gain comes at the expense of high-risk trades, whose costs rise, so state alignment pays off only for a venue that weighs low-risk costs more than 1.38 times as heavily as high-risk costs. Our code is available at https://github.com/moderncavemann/state-aligned-order-flow-contracts.

cs.AI↗

arXiv2026

MemTrial: Learning When to Trust Memory in LLM Portfolio Agents

Large language model (LLM) agents for portfolio management learn from experience: they credit each experience in their memory with the outcome of the decisions that used it. In financial markets, however, this outcome mostly reflects the market move shared by all decisions on that date, so the credit tracks the market rather than the experience, and these agents often do worse than simply holding the equal-weight (1/$N$) portfolio. We ask how an agent can credit an experience with what it changes, and answer it by putting memory on trial: drafts of the same decision with and without an experience face the same market, so the outcome they share cancels in their difference. Our agent, MemTrial, drafts each decision with eight combinations of its retrieved experiences, chosen by a fractional factorial design, and credits each experience with its Banzhaf value, the average of these differences. As each date occurs once and each draft is a noisy LLM sample, these credits are noisy and may not hold on new dates. MemTrial therefore pools them across dates and similar experiences with a hierarchical Bayesian model, acts on them only after they have predicted unseen dates, and otherwise stays anchored at a conservative reference such as 1/$N$. On four benchmarks, MemTrial not only benefits from experiences that matter (the best of 15 methods on a semi-synthetic benchmark with known experience quality) but also limits its losses when its values do not hold (at most 2.2\% below 1/$N$ on PortBench and InvestorBench, against 15--38\% for the best experience-learning agent). Averaged over five settings, it improves the utility of the best experience-learning agent by 21.2\%, and with eight LLMs it beats every LLM-based baseline on InvestorBench.

cs.AI↗

arXiv2026

Generate What You Can Trust: Content Credibility in Generative Recommenders

Generative recommendation (GR) represents items with semantic IDs (i.e., discrete token sequences) and generates target item tokens as recommendations. Despite its promising results, existing methods predominantly optimize for accuracy while neglecting the credibility of the recommendations they generate. This oversight inevitably exposes users to uncredible content (e.g., fake news) with serious societal consequences, including user distrust, reputation harm to platforms, and broader social instability. To address this critical yet underexplored challenge, we propose CreGR, the first credible GR model that jointly tackles content credibility across the two core stages of GR: tokenization and generation. In the tokenization stage, we design a new credibility-aware tokenizer that explicitly encourages the model to learn discriminative tokens respectively for credible and uncredible items, thereby disentangling credibility signals at the token level. Building on this, in the generation stage, we propose a novel accuracy-preserving and credibility-oriented generator grounded in discrete diffusion. Specifically, we introduce an asymmetric masking probability reduction strategy that selectively diminishes the contribution of tokens associated with uncredible content to the generation process, while leaving tokens encoding user preference signals unaffected so as to preserve recommendation accuracy. Experiments on three real-world datasets demonstrate the effectiveness of CreGR.

cs.IR↗

arXiv2026

SPRINT: Single-Step Generative Recommendation via Average Probability Velocity

Semantic ID (SID) based generative recommendation represents each item as a sequence of discrete tokens, and recommends by generating the SID of the item a user would like to interact with. Both dominant paradigms in this domain generally pay for generation token by token: autoregressive models decode the tokens left-to-right, while non-autoregressive models decode in parallel yet still need multiple rounds of refinement to stay competitive. Therefore, both generally spend multiple forward passes per item, a cost that is prohibitive in latency-sensitive recommender systems. We ask whether an item can be generated in a single forward pass, and answer it through a new perspective which we call average probability velocity. We view SID generation as a flow of token generation probabilities and characterize it by its average velocity over the whole generation process. We prove that this average velocity is fully determined by the average generation probability of each token. Therefore, we directly parameterize and learn the probabilities of all tokens in a single forward pass with a bidirectional Transformer. As these probabilities are generated independently across positions and the coherence among tokens is lost, we further design a dual-level flow contrastive objective to restore the coherence among an item's tokens. It contrasts the target SID against negative SIDs at both the token and SID levels. The token level ranks the generation probabilities of the target tokens above those of negative SIDs, while the SID level scores the tokens of each SID as a whole item for capturing token coherence of each item. Extensive experiments show that our model not only generates recommendations far more efficiently ($8.39-10.04\times$ speedup over the second-fastest AR/NAR method) but also attains superior recommendation accuracy ($7.77\%$ average improvement over the second-best.

cs.IR↗

arXiv2025

MIRNet: Integrating Constrained Graph-Based Reasoning with Pre-training for Diagnostic Medical Imaging

Automated interpretation of medical images demands robust modeling of complex visual-semantic relationships while addressing annotation scarcity, label imbalance, and clinical plausibility constraints. We introduce MIRNet (Medical Image Reasoner Network), a novel framework that integrates self-supervised pre-training with constrained graph-based reasoning. Tongue image diagnosis is a particularly challenging domain that requires fine-grained visual and semantic understanding. Our approach leverages self-supervised masked autoencoder (MAE) to learn transferable visual representations from unlabeled data; employs graph attention networks (GAT) to model label correlations through expert-defined structured graphs; enforces clinical priors via constraint-aware optimization using KL divergence and regularization losses; and mitigates imbalance using asymmetric loss (ASL) and boosting ensembles. To address annotation scarcity, we also introduce TongueAtlas-4K, a comprehensive expert-curated benchmark comprising 4,000 images annotated with 22 diagnostic labels--representing the largest public dataset in tongue analysis. Validation shows our method achieves state-of-the-art performance. While optimized for tongue diagnosis, the framework readily generalizes to broader diagnostic medical imaging tasks.

cs.CV↗

arXiv2025

A Survey on Enhancing Causal Reasoning Ability of Large Language Models

Large language models (LLMs) have recently shown remarkable performance in language tasks and beyond. However, due to their limited inherent causal reasoning ability, LLMs still face challenges in handling tasks that require robust causal reasoning ability, such as health-care and economic analysis. As a result, a growing body of research has focused on enhancing the causal reasoning ability of LLMs. Despite the booming research, there lacks a survey to well review the challenges, progress and future directions in this area. To bridge this significant gap, we systematically review literature on how to strengthen LLMs' causal reasoning ability in this paper. We start from the introduction of background and motivations of this topic, followed by the summarisation of key challenges in this area. Thereafter, we propose a novel taxonomy to systematically categorise existing methods, together with detailed comparisons within and between classes of methods. Furthermore, we summarise existing benchmarks and evaluation metrics for assessing LLMs' causal reasoning ability. Finally, we outline future research directions for this emerging field, offering insights and inspiration to researchers and practitioners in the area.

cs.CL↗

arXiv2023

A Game-theoretic Approach for Provably-Uniform Random Number Generation in Decentralized Networks

Many protocols in distributed computing rely on a source of randomness, usually called a random beacon, both for their applicability and security. This is especially true for proof-of-stake blockchain protocols in which the next miner or set of miners have to be chosen randomly and each party's likelihood to be selected is in proportion to their stake in the cryptocurrency. Current random beacons used in proof-of-stake protocols, such as Ouroboros and Algorand, have two fundamental limitations: Either (i)~they rely on pseudorandomness, e.g.~assuming that the output of a hash function is uniform, which is a widely-used but unproven assumption, or (ii)~they generate their randomness using a distributed protocol in which several participants are required to submit random numbers which are then used in the generation of a final random result. However, in this case, there is no guarantee that the numbers provided by the parties are uniformly random and there is no incentive for the parties to honestly generate uniform randomness. Most random beacons have both limitations. In this thesis, we provide a protocol for distributed generation of randomness. Our protocol does not rely on pseudorandomness at all. Similar to some of the previous approaches, it uses random inputs by different participants to generate a final random result. However, the crucial difference is that we provide a game-theoretic guarantee showing that it is in everyone's best interest to submit uniform random numbers. Hence, our approach is the first to incentivize honest behavior instead of just assuming it. Moreover, the approach is trustless and generates unbiased random numbers. It is also tamper-proof and no party can change the output or affect its distribution. Finally, it is designed with modularity in mind and can be easily plugged into existing distributed protocols such as proof-of-stake blockchains.

cs.CR↗

arXiv2020

Pathological Visual Question Answering

Is it possible to develop an "AI Pathologist" to pass the board-certified examination of the American Board of Pathology (ABP)? To build such a system, three challenges need to be addressed. First, we need to create a visual question answering (VQA) dataset where the AI agent is presented with a pathology image together with a question and is asked to give the correct answer. Due to privacy concerns, pathology images are usually not publicly available. Besides, only well-trained pathologists can understand pathology images, but they barely have time to help create datasets for AI research. The second challenge is: since it is difficult to hire highly experienced pathologists to create pathology visual questions and answers, the resulting pathology VQA dataset may contain errors. Training pathology VQA models using these noisy or even erroneous data will lead to problematic models that cannot generalize well on unseen images. The third challenge is: the medical concepts and knowledge covered in pathology question-answer (QA) pairs are very diverse while the number of QA pairs available for modeling training is limited. How to learn effective representations of diverse medical concepts based on limited data is technically demanding. In this paper, we aim to address these three challenges. To our best knowledge, our work represents the first one addressing the pathology VQA problem. To deal with the issue that a publicly available pathology VQA dataset is lacking, we create PathVQA dataset. To address the second challenge, we propose a learning-by-ignoring approach. To address the third challenge, we propose to use cross-modal self-supervised learning. We perform experiments on our created PathVQA dataset and the results demonstrate the effectiveness of our proposed learning-by-ignoring method and cross-modal self-supervised learning methods.

cs.CV↗