Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
119 tokens/sec
GPT-4o
56 tokens/sec
Gemini 2.5 Pro Pro
43 tokens/sec
o3 Pro
6 tokens/sec
GPT-4.1 Pro
47 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

NLLG Quarterly arXiv Report 09/23: What are the most influential current AI Papers? (2312.05688v1)

Published 9 Dec 2023 in cs.DL, cs.AI, cs.CL, cs.CV, cs.CY, and cs.LG

Abstract: AI has witnessed rapid growth, especially in the subfields NLP, Machine Learning (ML) and Computer Vision (CV). Keeping pace with this rapid progress poses a considerable challenge for researchers and professionals in the field. In this arXiv report, the second of its kind, which covers the period from January to September 2023, we aim to provide insights and analysis that help navigate these dynamic areas of AI. We accomplish this by 1) identifying the top-40 most cited papers from arXiv in the given period, comparing the current top-40 papers to the previous report, which covered the period January to June; 2) analyzing dataset characteristics and keyword popularity; 3) examining the global sectoral distribution of institutions to reveal differences in engagement across geographical areas. Our findings highlight the continued dominance of NLP: while only 16% of all submitted papers have NLP as primary category (more than 25% have CV and ML as primary category), 50% of the most cited papers have NLP as primary category, 90% of which target LLMs. Additionally, we show that i) the US dominates among both top-40 and top-9k papers, followed by China; ii) Europe clearly lags behind and is hardly represented in the top-40 most cited papers; iii) US industry is largely overrepresented in the top-40 most influential papers.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (43)
  1. Citations, citation indicators, and research quality: An overview of basic concepts and theories. Sage Open, 9(1):2158244019829575, 2019.
  2. Palm 2 technical report. arXiv preprint arXiv:2305.10403, 2023.
  3. A multitask, multilingual, multimodal evaluation of chatgpt on reasoning, hallucination, and interactivity, 2023.
  4. Rt-2: Vision-language-action models transfer web knowledge to robotic control. arXiv preprint arXiv:2307.15818, 2023.
  5. Sparks of artificial general intelligence: Early experiments with gpt-4. arXiv preprint arXiv:2303.12712, 2023.
  6. A survey on evaluation of large language models. arXiv preprint arXiv:2307.03109, 2023.
  7. Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality. See https://vicuna. lmsys. org (accessed 14 April 2023), 2023.
  8. Michael R Douglas. Large language models. arXiv preprint arXiv:2307.05782, 2023.
  9. Palm-e: An embodied multimodal language model. arXiv preprint arXiv:2303.03378, 2023.
  10. Nllg quarterly arxiv report 06/23: What are the most influential current ai papers? arXiv preprint arXiv:2308.04889, 2023.
  11. Imagebind: One embedding space to bind them all. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 15180–15190, 2023.
  12. Thresholds for statistical and clinical significance in systematic reviews with meta-analytic methods. BMC medical research methodology, 14(1):1–13, 2014.
  13. Neal Krause. Social support, stress, and well-being among older adults. Journal of gerontology, 41(4):512–519, 1986.
  14. Otter: A multi-modal model with in-context instruction tuning. arXiv preprint arXiv:2305.03726, 2023.
  15. Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models. arXiv preprint arXiv:2301.12597, 2023.
  16. Starcoder: may the source be with you! arXiv preprint arXiv:2305.06161, 2023.
  17. Let’s verify step by step. arXiv preprint arXiv:2305.20050, 2023.
  18. Lost in the middle: How language models use long contexts. arXiv preprint arXiv:2307.03172, 2023.
  19. Patrice Lopez. Grobid: Combining automatic bibliographic data recognition and term extraction for scholarship publications. In Research and Advanced Technology for Digital Libraries: 13th European Conference, ECDL 2009, Corfu, Greece, September 27-October 2, 2009. Proceedings 13, pages 473–474. Springer, 2009.
  20. Mark EJ Newman. Prediction of highly cited papers. Europhysics Letters, 105(2):28002, 2014.
  21. OpenAI. Gpt-4 technical report, 2023.
  22. Dinov2: Learning robust visual features without supervision. arXiv preprint arXiv:2304.07193, 2023.
  23. The refinedweb dataset for falcon llm: outperforming curated corpora with web data, and web data only. arXiv preprint arXiv:2306.01116, 2023.
  24. Savitzky-golay smoothing filters. Computers in Physics, 4(6):669–672, 1990.
  25. Direct preference optimization: Your language model is secretly a reward model. arXiv preprint arXiv:2305.18290, 2023.
  26. Toolformer: Language models can teach themselves to use tools, 2023.
  27. Norbert Schmitt. Quantifying word association responses: What is native-like? System, 26(3):389–401, 1998.
  28. Llama: Open and efficient foundation language models, 2023.
  29. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288, 2023.
  30. Visual chatgpt: Talking, drawing and editing with visual foundation models. arXiv preprint arXiv:2303.04671, 2023.
  31. The rise and potential of large language model based agents: A survey. arXiv preprint arXiv:2309.07864, 2023.
  32. Wizardlm: Empowering large language models to follow complex instructions. arXiv preprint arXiv:2304.12244, 2023.
  33. Baichuan 2: Open large-scale language models. arXiv preprint arXiv:2309.10305, 2023.
  34. Large language models as optimizers. arXiv preprint arXiv:2309.03409, 2023.
  35. Tree of thoughts: Deliberate problem solving with large language models. arXiv preprint arXiv:2305.10601, 2023.
  36. Adding conditional control to text-to-image diffusion models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 3836–3847, 2023.
  37. Siren’s song in the ai ocean: A survey on hallucination in large language models. arXiv preprint arXiv:2309.01219, 2023.
  38. A survey of large language models. arXiv preprint arXiv:2303.18223, 2023.
  39. Judging llm-as-a-judge with mt-bench and chatbot arena. arXiv preprint arXiv:2306.05685, 2023.
  40. A comprehensive survey on pretrained foundation models: A history from bert to chatgpt. arXiv preprint arXiv:2302.09419, 2023.
  41. Minigpt-4: Enhancing vision-language understanding with advanced large language models. arXiv preprint arXiv:2304.10592, 2023.
  42. Can large language models transform computational social science? arXiv preprint arXiv:2305.03514, 2023.
  43. Universal and transferable adversarial attacks on aligned language models. arXiv preprint arXiv:2307.15043, 2023.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (8)
  1. Ran Zhang (89 papers)
  2. Aida Kostikova (5 papers)
  3. Christoph Leiter (13 papers)
  4. Jonas Belouadi (12 papers)
  5. Daniil Larionov (12 papers)
  6. Yanran Chen (12 papers)
  7. Vivian Fresen (3 papers)
  8. Steffen Eger (90 papers)

Summary

Analysis of Influential AI Papers: An Examination of Trends and Contributions from the NLLG Quarterly arXiv Report

The paper "NLLG Quarterly arXiv Report 09/23" authored by Zhang et al. provides a rigorous analysis of the most influential AI papers from January to September 2023. The paper, which is based on arXiv submissions, offers a curated perspective on trends within subfields like NLP, Machine Learning (ML), and Computer Vision (CV). This report, an update from the mid-year analysis conducted in June 2023, offers several key insights into scholarly activity, trends, geographic distribution, and institutional contributions within AI.

Methodology and Data Collection

The authors employ a standardized methodology for data collection and analysis. They extracted papers from arXiv using categories such as cs.CL (Computation and Language), cs.LG (Machine Learning), cs.CV (Computer Vision and Pattern Recognition), and cs.AI (Artificial Intelligence). Papers from January to September 2023 were filtered based on these categories. Citation counts for these papers were fetched from Semantic Scholar, and these counts were normalized using z-scores to ensure fair comparison across different publication timelines.

The report builds on the methodology from the previous quarter, incorporating a new approach to compute stable z-scores by averaging z-scores computed across different week divisions. This robust statistical technique mitigates the dependency on specific week definitions, providing a more reliable measure of citation impact.

Key Findings

Dominance of NLP and LLM Topics

The analysis underscores the dominance of NLP and LLMs. While NLP papers constitute a smaller fraction of total submissions (16%), they hold a significant portion of the top-40 most cited papers (50%), with 90% of these centered on LLMs. This indicates a sustained interest and activity in NLP research, particularly in the development and evaluation of LLMs.

Geographic and Institutional Distribution

Geographically, the United States leads both in the volume of papers and citation impact, followed by China. This dominance is apparent in both academia and industry, although the industry's role is more pronounced. European contributions, by contrast, are sparse among the top-40 most cited papers, indicating a potential disparity in research output and impact.

The institutional analysis reveals that US-based companies such as Google, Meta, and OpenAI are primary contributors to the top-40 list. These organizations undertake significant independent research, demonstrated by high fractional scores, indicating substantial internal contributions without external collaborations. Conversely, academic institutions tend to engage more in collaborative research, reflecting a broader trend of cross-institutional partnerships.

Trends in AI Research

The keyword analysis highlights critical trends in AI research, with terms like "LLM," "GPT," "LLaMA," and "multimodality" showing increasing relevance. LLaMA models, for instance, have quickly risen in prominence following their introduction, challenging the long-standing dominance of ChatGPT-focused research.

This shift towards diverse LLM models and multimodality suggests a broadening of research horizons. There is an increasing exploration of applications beyond text generation, incorporating advancements in vision-language integration and other multimodal AI systems.

Implications and Future Directions

The findings have several implications for the AI research community. The robust growth in NLP and LLM-focused research underscores the necessity for continuous evaluation and benchmarking frameworks to assess these models' performance and ethical considerations. Additionally, the enhanced role of industry players signifies a collaborative potential that could further spur innovation through partnerships between academia and industry.

Going forward, the report suggests an ongoing need to monitor these dynamic trends. Regular updates and expanded datasets could provide even deeper insights into the evolving landscape of AI research, helping scholars and practitioners stay abreast of key developments and emerging topics.

In sum, the "NLLG Quarterly arXiv Report 09/23" offers a comprehensive view of influential AI research trends, highlighting the critical roles of NLP, LLMs, and multimodality. The report’s detailed analysis provides valuable perspectives on the contributions of various geographic regions and institutions, offering a roadmap for future research and collaborations in the rapidly evolving field of AI.

File Document Download Save Streamline Icon: https://streamlinehq.com PDF File Document Download Save Streamline Icon: https://streamlinehq.com Markdown
Youtube Logo Streamline Icon: https://streamlinehq.com

YouTube