Pith. sign in

REVIEW 24 cited by

Dynamic Prompt Learning via Policy Gradient for Semi-structured Mathematical Reasoning

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2209.14610 v3 pith:QXSZ2OGQ submitted 2022-09-29 cs.LG cs.AIcs.CLcs.CV

Pan Lu , Liang Qiu , Kai-Wei Chang , Ying Nian Wu , Song-Chun Zhu , Tanmay Rajpurohit , Peter Clark , Ashwin Kalyan This is my paper · ORCID
classification cs.LGcs.AIcs.CLcs.CV
keywords reasoningproblemsmathematicaltabmwptabulardataexamplesgpt-3
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Mathematical reasoning, a core ability of human intelligence, presents unique challenges for machines in abstract thinking and logical reasoning. Recent large pre-trained language models such as GPT-3 have achieved remarkable progress on mathematical reasoning tasks written in text form, such as math word problems (MWP). However, it is unknown if the models can handle more complex problems that involve math reasoning over heterogeneous information, such as tabular data. To fill the gap, we present Tabular Math Word Problems (TabMWP), a new dataset containing 38,431 open-domain grade-level problems that require mathematical reasoning on both textual and tabular data. Each question in TabMWP is aligned with a tabular context, which is presented as an image, semi-structured text, and a structured table. There are two types of questions: free-text and multi-choice, and each problem is annotated with gold solutions to reveal the multi-step reasoning process. We evaluate different pre-trained models on TabMWP, including the GPT-3 model in a few-shot setting. As earlier studies suggest, since few-shot GPT-3 relies on the selection of in-context examples, its performance is unstable and can degrade to near chance. The unstable issue is more severe when handling complex problems like TabMWP. To mitigate this, we further propose a novel approach, PromptPG, which utilizes policy gradient to learn to select in-context examples from a small amount of training data and then constructs the corresponding prompt for the test example. Experimental results show that our method outperforms the best baseline by 5.31% on the accuracy metric and reduces the prediction variance significantly compared to random selection, which verifies its effectiveness in selecting in-context examples.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 24 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Self-Consistency from Only Two Samples: CoT-PoT Ensembling for Efficient LLM Reasoning

    cs.CL 2026-04 unverdicted novelty 7.0 of 10

    Cross-modal agreement between chain-of-thought and program-of-thought reasoning enables self-consistency with only two LLM samples, reducing sampling cost by 9.3x while improving accuracy.

  2. Point-RFT: Improving Multimodal Reasoning with Visually Grounded Reinforcement Finetuning

    cs.CV 2025-05 reject novelty 7.0 of 10

    Point-RFT uses point-grounded chain-of-thought with GRPO reinforcement to improve chart reasoning, reporting 90.04% on ChartQA, though internal tables and the OOD setup weaken the claim.

  3. TaskGalaxy: Scaling Multi-modal Instruction Fine-tuning with Tens of Thousands Vision Task Types

    cs.CV 2025-02 conditional novelty 7.0 of 10

    The authors release TaskGalaxy, a 413k-sample multimodal instruction dataset with 19,227 hierarchically organized task types, and report consistent benchmark gains when it is added to LLaVA-v1.5 and InternVL-Chat-v1.0.

  4. MMAT-1M: A Large Reasoning Dataset for Multimodal Agent Tuning

    cs.CV 2025-07 conditional novelty 6.0 of 10

    A new one-million-sample multimodal agent tuning dataset with GPT-4o-generated rationales, reflection, and tool/RAG calls is shown to improve fine-tuned models, though training/eval benchmark overlap is not addressed.

  5. MarginSel : Max-Margin Demonstration Selection for LLMs

    cs.LG 2025-06 conditional novelty 6.0 of 10

    MarginSel picks demonstrations by matching the LLM's own ambiguous candidate-label sets between training and test instances, and reports consistent few-shot F1 gains over random and kNN baselines.

  6. CoMemo: LVLMs Need Image Context with Image Memory

    cs.CV 2025-06 conditional novelty 6.0 of 10

    CoMemo adds a cross-attention image-memory path and thumbnail-anchored position encoding to reduce visual neglect in long-context and multi-image LVLM tasks.

  7. What Really Matters for Table LLMs? A Meta-Evaluation of Model and Data Effects

    cs.CL 2025-01 conditional novelty 6.0 of 10

    By independently varying base models and training data across 12 models, this study shows that base model choice influences out-of-domain table task performance more than the instruction-tuning dataset does.

  8. InternLM-XComposer2.5-Reward: A Simple Yet Effective Multi-Modal Reward Model

    cs.CV 2025-01 conditional novelty 6.0 of 10

    IXC-2.5-Reward is an open-source multimodal reward model that achieves 70.0% macro accuracy on VL-RewardBench and improves LVLM chat via PPO.

  9. Eagle 2: Building Post-Training Data Strategies from Scratch for Frontier Vision-Language Models

    cs.CV 2025-01 conditional novelty 6.0 of 10

    Eagle2-9B matches or outperforms much larger vision-language models on many benchmarks through a carefully constructed post-training data strategy.

  10. HoVLE: Unleashing the Power of Monolithic Vision-Language Models with Holistic Vision-Language Embedding

    cs.CV 2024-12 conditional novelty 6.0 of 10

    HoVLE is a monolithic VLM whose holistic embedding module maps images and text into one shared space, letting a frozen LLM reach near-compositional performance.

  11. PVC: Progressive Visual Token Compression for Unified Image and Video Processing in Large Vision-Language Models

    cs.CV 2024-12 conditional novelty 6.0 of 10

    PVC unifies image and video token compression in VLMs by repeating images as static videos and using causal temporal attention with adaptive compression, achieving strong benchmark results at 64 tokens per frame.

  12. Mono-InternVL-1.5: Towards Cheaper and Faster Monolithic Multimodal Large Language Models

    cs.CV 2025-07 conditional novelty 5.5 of 10

    A monolithic multimodal LLM that cuts pre-training data by 58% and first-token latency by up to 69% while matching or beating its predecessor on 15 benchmarks.

  13. PaLMR: Towards Faithful Visual Reasoning via Multimodal Process Alignment

    cs.CV 2026-02 conditional novelty 5.0 of 10

    By gating the reinforcement-learning reward on a binary visual-fidelity score, PaLMR lifts Qwen2.5-VL-7B to 70.9 on HallusionBench and 59.3 on MMMU, surpassing its GRPO baseline.

  14. TableMind: An Autonomous Programmatic Agent for Tool-Augmented Table Reasoning

    cs.AI 2025-09 conditional novelty 5.0 of 10

    TableMind, a two-stage SFT-plus-RL agent trained on an 8B model, reports state-of-the-art results on three table reasoning benchmarks.

  15. A Compute-Matched Re-Evaluation of TroVE on MATH

    cs.PL 2025-07 conditional novelty 5.0 of 10

    After matching computational budget, TroVE's toolbox mechanism yields only a marginal, statistically non-significant 1% accuracy gain over a plain sampling baseline on MATH.

  16. Enhancing Chain-of-Thought Reasoning with Critical Representation Fine-tuning

    cs.CL 2025-07 conditional novelty 5.0 of 10

    CRFT selects critical internal representations via attention and saliency scores and fine-tunes only them, improving GSM8K accuracy over ReFT from 29.0% to 32.8% on LLaMA-2-7B.

  17. FinLMM-R1: Enhancing Financial Reasoning in LMM through Scalable Data and Reward Design

    cs.CL 2025-06 conditional novelty 5.0 of 10

    A two-stage RL framework with length, image-selection, and adversarial rewards, trained on 89,378 ASP-built financial image-question pairs, improves multimodal reasoning over LMM-R1.

  18. Multimodal Tabular Reasoning with Privileged Structured Information

    cs.LG 2025-06 conditional novelty 5.0 of 10

    An 8B multimodal LLM trained on 9k reasoning traces distilled from structured tables reaches state-of-the-art open-source accuracy on table-image question answering and fact verification.

  19. Advancing Multimodal Reasoning via Reinforcement Learning with Cold Start

    cs.CL 2025-05 conditional novelty 5.0 of 10

    Using a two-stage SFT-plus-GRPO pipeline, the authors train 3B and 7B multimodal models that beat prior open-source models on four math reasoning benchmarks, while showing that reflective 'aha moment' text is not a re...

  20. MMTABREAL: Real-World Benchmark for Multimodal Table Understanding

    cs.CV 2025-05 conditional novelty 5.0 of 10

    The paper releases a 500-table, 4,021-question benchmark of real-world multimodal tables and shows that leading vision-language models drop 20-40% in accuracy relative to earlier synthetic benchmarks.

  21. Texts or Images? A Fine-grained Analysis on the Effectiveness of Input Representations and Models for Table Question Answering

    cs.CL 2025-05 conditional novelty 5.0 of 10

    A controlled study shows that for small models, the best table representation depends on table size and question complexity, and the proposed FRES selection rule improves accuracy by about 10 points on average.

  22. CDW-CoT: Clustered Distance-Weighted Chain-of-Thoughts Reasoning

    cs.LG 2025-01 reject novelty 5.0 of 10

    CDW-CoT groups a reasoning dataset into clusters, learns a prompt distribution per cluster, and interpolates these distributions by embedding distance for each new query, reporting higher exact-match accuracy than thr...

  23. From Meta-Thought to Execution: Cognitively Aligned Post-Training for Generalizable and Reliable LLM Reasoning

    cs.AI 2026-01 conditional novelty 4.0 of 10

    Post-training LLMs first on abstract number-free reasoning plans (CoMT), then with confidence-weighted rewards (CCRL), raises math accuracy by ~2-5 points over standard CoT-SFT+RL across four models.

  24. Valley2: Exploring Multimodal Models with Scalable Vision-Language Design

    cs.CV 2025-01 conditional novelty 4.0 of 10

    Valley2, a 7B-scale open-source multimodal model, reports second-best OpenCompass average (67.4) among sub-10B models and the highest score (79.66) on its own in-house Ecom-VQA benchmark.

Pith tools