arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2606.11642cs.HCcs.CL
收藏

3-Key-Input: 探索文本输入的理论最少按键数

3-Key-Input: Exploring the Theoretical Minimum Keys for Text Entry

Naoki Kimura(金仓大学)

Naoki Kimura

AI总结:

本研究通过结合语言模型与2-5个物理按键,系统评估文本输入系统,发现3键+GPT-4o可实现字符错误率9.46%,表明在强语言模型先验下3键是实用最小值。

摘要:

如果我们为模糊键盘配备现代语言模型,可以将物理按键数量减少到多少?更少的按键在辅助设备和移动设备等受限场景中增加了硬件设计自由度。本文系统评估了使用2-5个物理按键结合基于语言模型的消歧的文本输入系统。在包含300个句子的英文语料库(商务/会话/技术各100句)上,我们比较了按键数量(2-5)、字母到按键映射(基于布局/基于频率/故意最坏情况)和解码器(仅Trie、GPT-2束搜索、GPT-4o选择)。我们发现,3键+GPT-4o实现了字符错误率(CER)9.46%和词错误率(WER)12.20%,相对于2键(CER 23.3%)CER降低了59%。在3键时,键流熵为1.54比特/字符;虽然增加到5键提高了准确率(CER 5.4%),但边际增益递减。在标准设计下,映射选择影响较小(ΔCER < 0.5个百分点),即使故意最坏映射也仅使CER增加+0.5个百分点,而技术句子的错误率大约是商务句子的两倍。这些结果表明,在我们评估的离线设置下,在强语言模型先验下,3键是通用英语的实用最小值。

英文摘要:

How far can we reduce the number of physical keys if we endow an ambiguous keyboard with modern language models? Fewer keys increase hardware design freedom in constrained settings such as assistive devices and mobile form factors. This paper systematically evaluates text entry systems using 2-5 physical keys combined with language-model-based disambiguation. On a 300-sentence English corpus (100 sentences each for Business / Conversational / Technical), we compare key counts (2-5), letter-to-key mappings (layout-based / frequency-based / intentionally worst-case), and decoders (Trie-only, GPT-2 beam search, GPT-4o selection). We find that 3 keys + GPT-4o achieves character error rate (CER) 9.46% and word error rate (WER) 12.20%, reducing CER by 59% relative to 2 keys (CER 23.3%). At 3 keys, the key-stream entropy is 1.54 bits/char; while increasing to 5 keys improves accuracy (CER 5.4%), the marginal gains diminish. Mapping choice has a small impact under standard designs (ΔCER < 0.5 pp), and even an intentionally worst mapping degrades CER by only +0.5 pp, whereas Technical sentences yield roughly twice the error rate of Business. These results suggest that, in our evaluated offline setting under a strong LM prior, 3 keys are a practical minimum for general English.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑