Pith. sign in

REVIEW 7 cited by

Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answering

Not yet reviewed by Pith; the record is open.

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 2209.09513 v2 pith:MAPEPRCG submitted 2022-09-20 cs.CL cs.AIcs.CVcs.LGcs.MM

classification cs.CLcs.AIcs.CVcs.LGcs.MM
keywords modelsansweringquestionexplanationslanguagesciencescienceqadata
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

When answering a question, humans utilize the information available across different modalities to synthesize a consistent and complete chain of thought (CoT). This process is normally a black box in the case of deep learning models like large-scale language models. Recently, science question benchmarks have been used to diagnose the multi-hop reasoning ability and interpretability of an AI system. However, existing datasets fail to provide annotations for the answers, or are restricted to the textual-only modality, small scales, and limited domain diversity. To this end, we present Science Question Answering (ScienceQA), a new benchmark that consists of ~21k multimodal multiple choice questions with a diverse set of science topics and annotations of their answers with corresponding lectures and explanations. We further design language models to learn to generate lectures and explanations as the chain of thought (CoT) to mimic the multi-hop reasoning process when answering ScienceQA questions. ScienceQA demonstrates the utility of CoT in language models, as CoT improves the question answering performance by 1.20% in few-shot GPT-3 and 3.99% in fine-tuned UnifiedQA. We also explore the upper bound for models to leverage explanations by feeding those in the input; we observe that it improves the few-shot performance of GPT-3 by 18.96%. Our analysis further shows that language models, similar to humans, benefit from explanations to learn from fewer data and achieve the same performance with just 40% of the data. The data and code are available at https://scienceqa.github.io.

Discussion (0). Sign in to comment.

Forward citations

Cited by 7 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. OpenAlex reports about 214 citations worldwide. Full citation record

  1. SlimVLM: Sensitivity-aware Dynamic Structured Pruning with Adaptive Visual Token Selection for Efficient Vision-Language Models

    cs.CV 2026-08 conditional novelty 6.0 of 10

    Visual-token attention filtering improves structured pruning of vision-language models, keeping 94% of average benchmark accuracy after removing 20% of parameters.

  2. SafeBuild-Bench: A Temporal-Robust Construction Safety Benchmark with Graph-Enhanced Data Mining

    cs.CV 2026-07 conditional novelty 6.0 of 10

    A temporal-aware construction-safety benchmark built from real inspection archives shows current multimodal models remain far from reliable, with best overall scores near 60.

  3. Spend Experts Where You Are Unsure: Confidence-Adaptive Routing for Mixture-of-Experts LoRA

    cs.LG 2026-07 conditional novelty 6.0 of 10

    Replacing fixed top-k routing in MoE-LoRA with router-confidence-based nucleus admission plus an expert-disagreement extension improves accuracy and OOD detection at matched average compute.

  4. TrustCLIP: Learning Private Visual Features via Adversarial Reconstruction

    cs.CV 2026-07 conditional novelty 6.0 of 10

    A lightweight adversarially trained projection degrades generative inversion of CLIP features while preserving most classification and VLM utility.

  5. Beyond Reasoning Gains: Mitigating General-Capability Forgetting in Large Reasoning Models

    cs.LG 2025-10 conditional novelty 6.0 of 10

    A dynamic replay and reweighting scheduler (RECAP) preserves general capabilities during RLVR while keeping reasoning performance at least as good as reasoning-only finetuning.

  6. AutoEvoEval: An Automated Framework for Evolving Close-Ended LLM Evaluation Data

    cs.CL 2025-06 conditional novelty 5.0 of 10

    AutoEvoEval applies 22 atomic perturbations and multi-round chains to MCQ benchmarks, causing average accuracy drops of 7.283% and up to 52.932% for long chains.

  7. A Comprehensive Survey of Deep Research: Systems, Methodologies, and Applications

    cs.AI 2025-06 conditional novelty 4.0 of 10

    A survey of 80+ Deep Research systems that proposes a four-layer taxonomy (foundation models, tool use, planning, synthesis) and compares commercial and open-source implementations.

Pith tools