{"url":"/method/ctc-loss","slug":"ctc-loss","name":"CTC Loss","full_name":"Connectionist Temporal Classification Loss","full_name_withheld":false,"description_markdown":"A **Connectionist Temporal Classification Loss**, or **CTC Loss**, is designed for tasks where we need alignment between sequences, but where that alignment is difficult - e.g. aligning each character to its location in an audio file. It calculates a loss between a continuous (unsegmented) time series and a target sequence. It does this by summing over the probability of possible alignments of input to target, producing a loss value which is differentiable with respect to each input node. The alignment of input to target is assumed to be “many-to-one”, which limits the length of the target sequence such that it must be $\\leq$ the input length.","description_state":"present","introduced_year":null,"introduced_by":{"title":null,"paper":null,"first_author":null,"n_authors":0,"url_abs":null,"archive_paper_url":null},"source":{"url":null,"title":null,"url_on_a_paper_host":false},"code_snippet_url":"https://github.com/pytorch/pytorch/blob/8850fd1952c3983793dcac4022fdc8e3913dad96/torch/nn/modules/loss.py#L1247","code_snippet_url_on_a_code_host":true,"categories":[{"area":"General","area_id":"general","collection":"Loss Functions","url":"/methods/category/loss-functions","pwc_aliases":[]}],"n_papers_tagged":47,"archive_num_papers":null,"papers_newest_first":[{"paper":null,"title":"Analyzing the Importance of Blank for CTC-Based Knowledge Distillation","date":"2025-06-02","arxiv_id":"2506.01503","n_code_links":0,"syntology":null},{"paper":null,"title":"Integrating Canonical Neural Units and Multi-Scale Training for Handwritten Text Recognition","date":"2024-10-24","arxiv_id":"2410.18374","n_code_links":0,"syntology":null},{"paper":null,"title":"Serialized Speech Information Guidance with Overlapped Encoding Separation for Multi-Speaker Automatic Speech Recognition","date":"2024-09-01","arxiv_id":"2409.00815","n_code_links":0,"syntology":null},{"paper":"/paper/li-tta-language-informed-test-time-adaptation","title":"LI-TTA: Language Informed Test-Time Adaptation for Automatic Speech Recognition","date":"2024-08-11","arxiv_id":"2408.05769","n_code_links":1,"syntology":null},{"paper":null,"title":"Improving Robustness of LLM-based Speech Synthesis by Learning Monotonic Alignment","date":"2024-06-25","arxiv_id":"2406.17957","n_code_links":0,"syntology":null},{"paper":"/paper/best-practices-for-a-handwritten-text","title":"Best Practices for a Handwritten Text Recognition System","date":"2024-04-17","arxiv_id":"2404.11339","n_code_links":1,"syntology":null},{"paper":null,"title":"AdaMER-CTC: Connectionist Temporal Classification with Adaptive Maximum Entropy Regularization for Automatic Speech Recognition","date":"2024-03-18","arxiv_id":"2403.11578","n_code_links":0,"syntology":null},{"paper":"/paper/key-frame-mechanism-for-efficient-conformer","title":"Key Frame Mechanism For Efficient Conformer Based End-to-end Speech Recognition","date":"2023-10-23","arxiv_id":"2310.14954","n_code_links":1,"syntology":null},{"paper":"/paper/self-distillation-regularized-connectionist","title":"Self-distillation Regularized Connectionist Temporal Classification Loss for Text Recognition: A Simple Yet Effective Approach","date":"2023-08-17","arxiv_id":"2308.08806","n_code_links":1,"syntology":null},{"paper":null,"title":"Alternative Pseudo-Labeling for Semi-Supervised Automatic Speech Recognition","date":"2023-08-12","arxiv_id":"2308.06547","n_code_links":0,"syntology":null},{"paper":null,"title":"Align With Purpose: Optimize Desired Properties in CTC Models with a General Plug-and-Play Framework","date":"2023-07-04","arxiv_id":"2307.01715","n_code_links":0,"syntology":null},{"paper":null,"title":"Improving Non-autoregressive Translation Quality with Pretrained Language Model, Embedding Distillation and Upsampling Strategy for CTC","date":"2023-06-10","arxiv_id":"2306.06345","n_code_links":0,"syntology":null},{"paper":null,"title":"INTapt: Information-Theoretic Adversarial Prompt Tuning for Enhanced Non-Native Speech Recognition","date":"2023-05-25","arxiv_id":"2305.16371","n_code_links":0,"syntology":null},{"paper":"/paper/softcorrect-error-correction-with-soft","title":"SoftCorrect: Error Correction with Soft Detection for Automatic Speech Recognition","date":"2022-12-02","arxiv_id":"2212.01039","n_code_links":1,"syntology":{"ran":6,"of":6,"unverified":0,"pointer_only":0}},{"paper":"/paper/weakly-supervised-fingerspelling-recognition","title":"Weakly-supervised Fingerspelling Recognition in British Sign Language Videos","date":"2022-11-16","arxiv_id":"2211.08954","n_code_links":1,"syntology":null},{"paper":null,"title":"Peak-First CTC: Reducing the Peak Latency of CTC Models by Applying Peak-First Regularization","date":"2022-11-07","arxiv_id":"2211.03284","n_code_links":0,"syntology":null},{"paper":"/paper/temporal-superimposed-crossover-module-for","title":"Temporal superimposed crossover module for effective continuous sign language","date":"2022-11-07","arxiv_id":"2211.03387","n_code_links":1,"syntology":null},{"paper":"/paper/trimtail-low-latency-streaming-asr-with","title":"TrimTail: Low-Latency Streaming ASR with Simple but Effective Spectrogram-Level Length Penalty","date":"2022-11-01","arxiv_id":"2211.00522","n_code_links":1,"syntology":{"ran":3,"of":5,"unverified":2,"pointer_only":0}},{"paper":null,"title":"Uconv-Conformer: High Reduction of Input Sequence Length for End-to-End Speech Recognition","date":"2022-08-16","arxiv_id":"2208.07657","n_code_links":0,"syntology":null},{"paper":null,"title":"Leveraging Acoustic Contextual Representation by Audio-textual Cross-modal Learning for Conversational ASR","date":"2022-07-03","arxiv_id":"2207.01039","n_code_links":0,"syntology":null},{"paper":null,"title":"Multi-scale temporal network for continuous sign language recognition","date":"2022-04-08","arxiv_id":"2204.03864","n_code_links":0,"syntology":null},{"paper":null,"title":"Non-Autoregressive ASR with Self-Conditioned Folded Encoders","date":"2022-02-17","arxiv_id":"2202.08474","n_code_links":0,"syntology":null},{"paper":null,"title":"PM-MMUT: Boosted Phone-Mask Data Augmentation using Multi-Modeling Unit Training for Phonetic-Reduction-Robust E2E Speech Recognition","date":"2021-12-13","arxiv_id":"2112.06721","n_code_links":0,"syntology":null},{"paper":"/paper/fast-md-fast-multi-decoder-end-to-end-speech","title":"Fast-MD: Fast Multi-Decoder End-to-End Speech Translation with Non-Autoregressive Hidden Intermediates","date":"2021-09-27","arxiv_id":"2109.12804","n_code_links":1,"syntology":null},{"paper":"/paper/golos-russian-dataset-for-speech-research","title":"Golos: Russian Dataset for Speech Research","date":"2021-06-18","arxiv_id":"2106.10161","n_code_links":2,"syntology":{"ran":4,"of":5,"unverified":1,"pointer_only":3}},{"paper":"/paper/multi-speaker-asr-combining-non","title":"Multi-Speaker ASR Combining Non-Autoregressive Conformer CTC and Conditional Speaker Chain","date":"2021-06-16","arxiv_id":"2106.08595","n_code_links":1,"syntology":null},{"paper":"/paper/why-does-ctc-result-in-peaky-behavior","title":"Why does CTC result in peaky behavior?","date":"2021-05-31","arxiv_id":"2105.14849","n_code_links":1,"syntology":null},{"paper":null,"title":"Relaxing the Conditional Independence Assumption of CTC-based ASR by Conditioning on Intermediate Predictions","date":"2021-04-06","arxiv_id":"2104.02724","n_code_links":0,"syntology":null},{"paper":null,"title":"Multiple-hypothesis CTC-based semi-supervised adaptation of end-to-end speech recognition","date":"2021-03-29","arxiv_id":"2103.15515","n_code_links":0,"syntology":null},{"paper":null,"title":"Intermediate Loss Regularization for CTC-based Speech Recognition","date":"2021-02-05","arxiv_id":"2102.03216","n_code_links":0,"syntology":null}],"papers_shown":30,"tasks":[{"task":"/task/speech-recognition","name":"Speech Recognition","papers":24},{"task":"/task/speech-recognition-1","name":"speech-recognition","papers":23},{"task":"/task/automatic-speech-recognition-2","name":"Automatic Speech Recognition","papers":18},{"task":"/task/automatic-speech-recognition","name":"Automatic Speech Recognition (ASR)","papers":18},{"task":"/task/language-modelling","name":"Language Modelling","papers":10},{"task":"/task/language-modeling","name":"Language Modeling","papers":8},{"task":"/task/decoder","name":"Decoder","papers":6},{"task":"/task/knowledge-distillation","name":"Knowledge Distillation","papers":4},{"task":"/task/handwritten-text-recognition","name":"Handwritten Text Recognition","papers":3},{"task":"/task/lipreading","name":"Lipreading","papers":3},{"task":"/task/sign-language-recognition","name":"Sign Language Recognition","papers":3},{"task":"/task/translation","name":"Translation","papers":3},{"task":"/task/audio-visual-speech-recognition","name":"Audio-Visual Speech Recognition","papers":2},{"task":null,"name":"CPU","papers":2},{"task":"/task/data-augmentation","name":"Data Augmentation","papers":2},{"task":null,"name":"GPU","papers":2},{"task":"/task/classification","name":"General Classification","papers":2},{"task":"/task/multi-task-learning","name":"Multi-Task Learning","papers":2},{"task":"/task/transfer-learning","name":"Transfer Learning","papers":2},{"task":"/task/visual-speech-recognition","name":"Visual Speech Recognition","papers":2}],"tasks_shown":20,"n_tasks":42,"usage_by_year":[{"year":"2016","papers":1},{"year":"2017","papers":2},{"year":"2018","papers":4},{"year":"2019","papers":4},{"year":"2020","papers":5},{"year":"2021","papers":9},{"year":"2022","papers":9},{"year":"2023","papers":6},{"year":"2024","papers":6},{"year":"2025","papers":1}],"row_source":"embedded","archive":{"source":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","archive_url":"https://paperswithcode.com/method/ctc-loss"},"syntology_read_at":"2026-09-24T18:15:14+00:00"}