{"work":{"id":"438f105c-fa9b-44aa-ad52-43acb8045cda","openalex_id":"https://openalex.org/W4409051416","doi":"10.48550/arxiv.2503.20215","arxiv_id":"2503.20215","raw_key":null,"title":"Qwen2.5-Omni Technical Report","authors":null,"authors_text":"Jin Xu, Zhifang Guo, Jinzheng He, Hangrui Hu, Ting He, Shuai Bai","year":2025,"venue":"cs.CL","abstract":"In this report, we present Qwen2.5-Omni, an end-to-end multimodal model designed to perceive diverse modalities, including text, images, audio, and video, while simultaneously generating text and natural speech responses in a streaming manner. To enable the streaming of multimodal information inputs, both audio and visual encoders utilize a block-wise processing approach. To synchronize the timestamps of video inputs with audio, we organize the audio and video sequentially in an interleaved manner and propose a novel position embedding approach, named TMRoPE(Time-aligned Multimodal RoPE). To concurrently generate text and speech while avoiding interference between the two modalities, we propose \\textbf{Thinker-Talker} architecture. In this framework, Thinker functions as a large language model tasked with text generation, while Talker is a dual-track autoregressive model that directly utilizes the hidden representations from the Thinker to produce audio tokens as output. Both the Thinker and Talker models are designed to be trained and inferred in an end-to-end manner. For decoding audio tokens in a streaming manner, we introduce a sliding-window DiT that restricts the receptive field, aiming to reduce the initial package delay. Qwen2.5-Omni is comparable with the similarly sized Qwen2.5-VL and outperforms Qwen2-Audio. Furthermore, Qwen2.5-Omni achieves state-of-the-art performance on multimodal benchmarks like Omni-Bench. Notably, Qwen2.5-Omni's performance in end-to-end speech instruction following is comparable to its capabilities with text inputs, as evidenced by benchmarks such as MMLU and GSM8K. As for speech generation, Qwen2.5-Omni's streaming Talker outperforms most existing streaming and non-streaming alternatives in robustness and naturalness.","external_url":"https://arxiv.org/abs/2503.20215","cited_by_count":8,"metadata_source":"pith","metadata_fetched_at":"2026-08-05T02:28:24.338817+00:00","pith_arxiv_id":"2503.20215","created_at":"2026-05-09T06:45:44.072529+00:00","updated_at":"2026-08-05T02:28:24.338817+00:00","title_quality_ok":false,"display_title":"Qwen2.5-Omni Technical Report","render_title":"Qwen2.5-Omni Technical Report"},"hub":{"state":{"work_id":"438f105c-fa9b-44aa-ad52-43acb8045cda","tier":"super_hub","tier_reason":"100+ Pith inbound or 10,000+ external citations","pith_inbound_count":225,"external_cited_by_count":8,"distinct_field_count":12,"first_pith_cited_at":"2024-11-26T18:57:29+00:00","last_pith_cited_at":"2026-07-07T17:43:36+00:00","author_build_status":"needed","summary_status":"needed","contexts_status":"needed","graph_status":"needed","ask_index_status":"needed","reader_status":"not_needed","recognition_status":"not_needed","updated_at":"2026-08-21T11:29:32.165405+00:00","tier_text":"super_hub"},"tier":"super_hub","role_counts":[{"context_role":"background","n":20},{"context_role":"baseline","n":10},{"context_role":"method","n":2}],"polarity_counts":[{"context_polarity":"background","n":19},{"context_polarity":"baseline","n":10},{"context_polarity":"use_method","n":2},{"context_polarity":"unclear","n":1}],"runs":{"ask_index":{"job_type":"ask_index","status":"succeeded","result":{"title":"Qwen2.5-Omni Technical Report","claims":[{"claim_text":"In this report, we present Qwen2.5-Omni, an end-to-end multimodal model designed to perceive diverse modalities, including text, images, audio, and video, while simultaneously generating text and natural speech responses in a streaming manner. To enable the streaming of multimodal information inputs, both audio and visual encoders utilize a block-wise processing approach. To synchronize the timestamps of video inputs with audio, we organize the audio and video sequentially in an interleaved manner and propose a novel position embedding approach, named TMRoPE(Time-aligned Multimodal RoPE). To c","claim_type":"abstract","evidence_strength":"source_metadata"},{"claim_text":"environments, and adaptive evaluation protocols tailored to different task paradigms. 1) Model Selection and Deployment Setup:We comprehensively evaluate our proposed FM-Speech against 11 advanced speech LLMs, comprising eight mainstream open-source models (Audio Flamingo 3 [8], Qwen3-Omni [7], Kimi-Audio [25], Step-Audio 2 [26], Omni- Captioner [27], Mimo-Audio [28], Qwen2.5-Omni [29], and Qwen2- Audio [30]) and three representative proprietary models (Gemini 2.5 Flash, Gemini 3 Flash, and Gemi","claim_type":"baseline","confidence":0.95,"evidence_strength":"citation_context"},{"claim_text":"a textual prompt as input, and generate text responses. Baselines.We compare GRM with five baselines from text transfer and native audio settings. The text transfer baselines are GCG [43] and AutoDAN [22], where adversarial prompts are gen- erated in text and converted to speech with gTTS. The native audio baselines include BoN [15], Jailbreak-AudioBench (AudioBench) [4], and SSJ [37]. BoN and AudioBench were originally evaluated with multi-variant sampling, whereas GRM generates a single advers","claim_type":"baseline","confidence":0.95,"evidence_strength":"citation_context"},{"claim_text":"[36,47] and an audio encoder [8] produce modality-specific embeddings, which are projected into a shared space and fed alongside tokenized text into an LLM backbone. Early AVLMs [42,54] demonstrated the viability of this paradigm, and more recent models [10,44,52] achieve competitive performance on benchmarks requiring joint audio-visual reasoning [27,31,53]. 4 Baid et al. 2.2 Hallucination in LLMs and VLMs Hallucination has been studied extensively in LLMs [17,19] and VLMs [15,29,38]. Training ","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"SLMs aim to enable seamless spoken dialogue interactions [ 1]. A typical SLM consists of three main components: 1) a speech encoder that converts the input waveform into speech representations, typically either continuous speech embeddings [ 22, 23] or discrete speech tokens [ 26-28]; 2) an LLM backbone that processes these representations and generates outputs, which may be text tokens [23], speech tokens [14], or both [12]; and 3) a speech decoder that synthesizes the model output into spoken ","claim_type":"baseline","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"driven remarkable breakthroughs in large language models (LLMs), empowering them with robust reasoning capabilities to solve intricate mathematical prob- lems [8,10,44] and generate high-quality, functional code [27,61]. Despite these significant advancements in purely textual domains, how to thoroughly explore and effectively enhance the reasoning capabilities of omni-modal models [41,42] within the context of integrated multi-modal processing remains an open and challenging problem [2,18,57]. ","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"depends on both perception and reasoning. In this work, we present the first systematic empirical study of uncertainty estimation in audio-aware LLMs. We evaluate several representative methods, including predictive entropy, length-normalized entropy, semantic entropy [65], discrete semantic entropy [65], and P(True) [62], across mul- tiple state-of-the-art ALLMs [9], [11] and a diverse set of benchmarks [26], [27], [29]-[31], [44]. Our evaluation spans both general audio understanding and reaso","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"}],"why_cited":"Pith tracks Qwen2.5-Omni Technical Report because it crossed a citation-hub threshold. Current citing contexts most often use it as background evidence (11 contexts).","role_counts":[{"n":11,"context_role":"background"},{"n":8,"context_role":"baseline"},{"n":2,"context_role":"method"}]},"error":null,"updated_at":"2026-05-18T05:20:34.388425+00:00"},"author_expand":{"job_type":"author_expand","status":"succeeded","result":{"authors_linked":[{"id":"c42757d0-ace5-480f-b243-890854f9054c","orcid":null,"display_name":"Jin Xu"},{"id":"6fe3c884-415c-45f2-b8fb-6629cd6c24df","orcid":null,"display_name":"Zhifang Guo"},{"id":"c6a51477-94c9-40f9-bdd9-97f195765b54","orcid":null,"display_name":"Jinzheng He"},{"id":"d939bd6b-83ae-4395-8119-a82cb1101e6d","orcid":null,"display_name":"Hangrui Hu"},{"id":"64156499-4cb1-4de8-9683-9d46746db822","orcid":null,"display_name":"Ting He"},{"id":"2bc8cfa9-c2ce-48ef-87d4-e0de52ae9a92","orcid":null,"display_name":"Shuai Bai"}]},"error":null,"updated_at":"2026-05-18T05:20:35.013711+00:00"},"context_extract":{"job_type":"context_extract","status":"succeeded","result":{"enqueued_papers":25},"error":null,"updated_at":"2026-05-14T08:47:58.300349+00:00"},"graph_features":{"job_type":"graph_features","status":"succeeded","result":{"co_cited":[{"title":"Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities","work_id":"008df105-2fdd-45d8-857a-8e35868aecb6","shared_citers":26},{"title":"Qwen3-Omni Technical Report","work_id":"ae43e594-8bab-4471-b6af-92a300f6a048","shared_citers":23},{"title":"Qwen2-Audio Technical Report","work_id":"c249e63c-cf40-408f-a4ff-fdf68e8cbeb8","shared_citers":21},{"title":"GPT-4o System Card","work_id":"f37bf1c7-4964-4e56-9762-d20da8d9009f","shared_citers":19},{"title":"Kimi-Audio Technical Report","work_id":"9c2ba56b-5585-4f28-b751-703f31dca2d5","shared_citers":16},{"title":"Qwen3 Technical Report","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","shared_citers":14},{"title":"VideoLLaMA 2: Advancing Spatial-Temporal Modeling and Audio Understanding in Video-LLMs","work_id":"ccfc3f89-c510-45f1-8a35-ed1a56c0ae5c","shared_citers":14},{"title":"MMAU: A Massive Multi-Task Audio Understanding and Reasoning Benchmark","work_id":"e60f85db-636c-4830-af85-5d31ebc74a1b","shared_citers":11},{"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","shared_citers":10},{"title":"Step-audio 2 technical report","work_id":"2317bc9f-2d58-4e53-b7ea-804337e9fdef","shared_citers":10},{"title":"arXiv preprint arXiv:2412.02612 , year=","work_id":"1d250ff4-6ca5-4eb5-b561-48106b630d8b","shared_citers":9},{"title":"DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning","work_id":"e6b75ad5-2877-4168-97c8-710407094d20","shared_citers":9},{"title":"Mmsu: A massive multi-task spoken language understanding and reasoning benchmark","work_id":"1a24d3b1-2d00-407c-90f0-b0aeec13bfe6","shared_citers":9},{"title":"Moshi: a speech-text foundation model for real-time dialogue","work_id":"3104332b-d279-44c8-aaa7-3d5a13c01832","shared_citers":9},{"title":"Audio flamingo 3: Advancing audio intelligence with fully open large audio language models","work_id":"67c2892d-8e27-4da1-8198-71c48e673e96","shared_citers":8},{"title":"Daily-omni: Towards audio-visual reasoning with temporal alignment across modalities","work_id":"8c7aec87-020f-4959-9199-7df8b9231cc4","shared_citers":8},{"title":"Phi-4-Mini Technical Report: Compact yet Powerful Multimodal Language Models via Mixture-of-LoRAs","work_id":"83956045-536a-41ff-af02-b80e2a614eab","shared_citers":8},{"title":"Qwen2.5-VL Technical Report","work_id":"69dffacb-bfe8-442d-be86-48624c60426f","shared_citers":8},{"title":"The Llama 3 Herd of Models","work_id":"1549a635-88af-4ac1-acfe-51ae7bb53345","shared_citers":8},{"title":"DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models","work_id":"c5006563-f3ec-438a-9e35-b7b484f34828","shared_citers":7},{"title":"LLaMA: Open and Efficient Foundation Language Models","work_id":"c018fc23-6f3f-4035-9d02-28a2173b2b9d","shared_citers":7},{"title":"Qwen3-VL Technical Report","work_id":"1fe243aa-e3c0-4da6-b391-4cbcfc88d5c0","shared_citers":7},{"title":"Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Models","work_id":"d3f033ac-bfa8-4143-9d0d-51f3f5bd3f0e","shared_citers":7},{"title":"Qwen Technical Report","work_id":"bb1fd52f-6b2f-437c-9516-37bdf6eb9be8","shared_citers":7}],"time_series":[{"n":1,"year":2025},{"n":67,"year":2026}],"dependency_candidates":[]},"error":null,"updated_at":"2026-05-14T08:58:01.053210+00:00"},"identity_refresh":{"job_type":"identity_refresh","status":"succeeded","result":{"items":[{"title":"Qwen3 Technical Report","outcome":"unchanged","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","resolver":"local_arxiv","confidence":0.98,"old_work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e"}],"counts":{"fixed":0,"merged":0,"unchanged":1,"quarantined":0,"needs_external_resolution":0},"errors":[],"attempted":1},"error":null,"updated_at":"2026-05-14T08:48:00.610696+00:00"},"role_polarity":{"job_type":"role_polarity","status":"succeeded","result":{"title":"Qwen2.5-Omni Technical Report","claims":[{"claim_text":"In this report, we present Qwen2.5-Omni, an end-to-end multimodal model designed to perceive diverse modalities, including text, images, audio, and video, while simultaneously generating text and natural speech responses in a streaming manner. To enable the streaming of multimodal information inputs, both audio and visual encoders utilize a block-wise processing approach. To synchronize the timestamps of video inputs with audio, we organize the audio and video sequentially in an interleaved manner and propose a novel position embedding approach, named TMRoPE(Time-aligned Multimodal RoPE). To c","claim_type":"abstract","evidence_strength":"source_metadata"},{"claim_text":"environments, and adaptive evaluation protocols tailored to different task paradigms. 1) Model Selection and Deployment Setup:We comprehensively evaluate our proposed FM-Speech against 11 advanced speech LLMs, comprising eight mainstream open-source models (Audio Flamingo 3 [8], Qwen3-Omni [7], Kimi-Audio [25], Step-Audio 2 [26], Omni- Captioner [27], Mimo-Audio [28], Qwen2.5-Omni [29], and Qwen2- Audio [30]) and three representative proprietary models (Gemini 2.5 Flash, Gemini 3 Flash, and Gemi","claim_type":"baseline","confidence":0.95,"evidence_strength":"citation_context"},{"claim_text":"a textual prompt as input, and generate text responses. Baselines.We compare GRM with five baselines from text transfer and native audio settings. The text transfer baselines are GCG [43] and AutoDAN [22], where adversarial prompts are gen- erated in text and converted to speech with gTTS. The native audio baselines include BoN [15], Jailbreak-AudioBench (AudioBench) [4], and SSJ [37]. BoN and AudioBench were originally evaluated with multi-variant sampling, whereas GRM generates a single advers","claim_type":"baseline","confidence":0.95,"evidence_strength":"citation_context"},{"claim_text":"[36,47] and an audio encoder [8] produce modality-specific embeddings, which are projected into a shared space and fed alongside tokenized text into an LLM backbone. Early AVLMs [42,54] demonstrated the viability of this paradigm, and more recent models [10,44,52] achieve competitive performance on benchmarks requiring joint audio-visual reasoning [27,31,53]. 4 Baid et al. 2.2 Hallucination in LLMs and VLMs Hallucination has been studied extensively in LLMs [17,19] and VLMs [15,29,38]. Training ","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"SLMs aim to enable seamless spoken dialogue interactions [ 1]. A typical SLM consists of three main components: 1) a speech encoder that converts the input waveform into speech representations, typically either continuous speech embeddings [ 22, 23] or discrete speech tokens [ 26-28]; 2) an LLM backbone that processes these representations and generates outputs, which may be text tokens [23], speech tokens [14], or both [12]; and 3) a speech decoder that synthesizes the model output into spoken ","claim_type":"baseline","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"driven remarkable breakthroughs in large language models (LLMs), empowering them with robust reasoning capabilities to solve intricate mathematical prob- lems [8,10,44] and generate high-quality, functional code [27,61]. Despite these significant advancements in purely textual domains, how to thoroughly explore and effectively enhance the reasoning capabilities of omni-modal models [41,42] within the context of integrated multi-modal processing remains an open and challenging problem [2,18,57]. ","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"},{"claim_text":"depends on both perception and reasoning. In this work, we present the first systematic empirical study of uncertainty estimation in audio-aware LLMs. We evaluate several representative methods, including predictive entropy, length-normalized entropy, semantic entropy [65], discrete semantic entropy [65], and P(True) [62], across mul- tiple state-of-the-art ALLMs [9], [11] and a diverse set of benchmarks [26], [27], [29]-[31], [44]. Our evaluation spans both general audio understanding and reaso","claim_type":"background","confidence":0.9,"evidence_strength":"citation_context"}],"why_cited":"Pith tracks Qwen2.5-Omni Technical Report because it crossed a citation-hub threshold. Current citing contexts most often use it as background evidence (11 contexts).","role_counts":[{"n":11,"context_role":"background"},{"n":8,"context_role":"baseline"},{"n":2,"context_role":"method"}]},"error":null,"updated_at":"2026-05-18T05:20:35.018109+00:00"},"summary_claims":{"job_type":"summary_claims","status":"succeeded","result":{"title":"Qwen2.5-Omni Technical Report","claims":[{"claim_text":"In this report, we present Qwen2.5-Omni, an end-to-end multimodal model designed to perceive diverse modalities, including text, images, audio, and video, while simultaneously generating text and natural speech responses in a streaming manner. To enable the streaming of multimodal information inputs, both audio and visual encoders utilize a block-wise processing approach. To synchronize the timestamps of video inputs with audio, we organize the audio and video sequentially in an interleaved manner and propose a novel position embedding approach, named TMRoPE(Time-aligned Multimodal RoPE). To c","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks Qwen2.5-Omni Technical Report because it crossed a citation-hub threshold.","role_counts":[]},"error":null,"updated_at":"2026-05-14T08:58:01.055588+00:00"}},"summary":{"title":"Qwen2.5-Omni Technical Report","claims":[{"claim_text":"In this report, we present Qwen2.5-Omni, an end-to-end multimodal model designed to perceive diverse modalities, including text, images, audio, and video, while simultaneously generating text and natural speech responses in a streaming manner. To enable the streaming of multimodal information inputs, both audio and visual encoders utilize a block-wise processing approach. To synchronize the timestamps of video inputs with audio, we organize the audio and video sequentially in an interleaved manner and propose a novel position embedding approach, named TMRoPE(Time-aligned Multimodal RoPE). To c","claim_type":"abstract","evidence_strength":"source_metadata"}],"why_cited":"Pith tracks Qwen2.5-Omni Technical Report because it crossed a citation-hub threshold.","role_counts":[]},"graph":{"co_cited":[{"title":"Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities","work_id":"008df105-2fdd-45d8-857a-8e35868aecb6","shared_citers":26},{"title":"Qwen3-Omni Technical Report","work_id":"ae43e594-8bab-4471-b6af-92a300f6a048","shared_citers":23},{"title":"Qwen2-Audio Technical Report","work_id":"c249e63c-cf40-408f-a4ff-fdf68e8cbeb8","shared_citers":21},{"title":"GPT-4o System Card","work_id":"f37bf1c7-4964-4e56-9762-d20da8d9009f","shared_citers":19},{"title":"Kimi-Audio Technical Report","work_id":"9c2ba56b-5585-4f28-b751-703f31dca2d5","shared_citers":16},{"title":"Qwen3 Technical Report","work_id":"25a4e30c-1232-48e7-9925-02fa12ba7c9e","shared_citers":14},{"title":"VideoLLaMA 2: Advancing Spatial-Temporal Modeling and Audio Understanding in Video-LLMs","work_id":"ccfc3f89-c510-45f1-8a35-ed1a56c0ae5c","shared_citers":14},{"title":"MMAU: A Massive Multi-Task Audio Understanding and Reasoning Benchmark","work_id":"e60f85db-636c-4830-af85-5d31ebc74a1b","shared_citers":11},{"title":"GPT-4 Technical Report","work_id":"b928e041-6991-4c08-8c81-0359e4097c7b","shared_citers":10},{"title":"Step-audio 2 technical report","work_id":"2317bc9f-2d58-4e53-b7ea-804337e9fdef","shared_citers":10},{"title":"arXiv preprint arXiv:2412.02612 , year=","work_id":"1d250ff4-6ca5-4eb5-b561-48106b630d8b","shared_citers":9},{"title":"DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning","work_id":"e6b75ad5-2877-4168-97c8-710407094d20","shared_citers":9},{"title":"Mmsu: A massive multi-task spoken language understanding and reasoning benchmark","work_id":"1a24d3b1-2d00-407c-90f0-b0aeec13bfe6","shared_citers":9},{"title":"Moshi: a speech-text foundation model for real-time dialogue","work_id":"3104332b-d279-44c8-aaa7-3d5a13c01832","shared_citers":9},{"title":"Audio flamingo 3: Advancing audio intelligence with fully open large audio language models","work_id":"67c2892d-8e27-4da1-8198-71c48e673e96","shared_citers":8},{"title":"Daily-omni: Towards audio-visual reasoning with temporal alignment across modalities","work_id":"8c7aec87-020f-4959-9199-7df8b9231cc4","shared_citers":8},{"title":"Phi-4-Mini Technical Report: Compact yet Powerful Multimodal Language Models via Mixture-of-LoRAs","work_id":"83956045-536a-41ff-af02-b80e2a614eab","shared_citers":8},{"title":"Qwen2.5-VL Technical Report","work_id":"69dffacb-bfe8-442d-be86-48624c60426f","shared_citers":8},{"title":"The Llama 3 Herd of Models","work_id":"1549a635-88af-4ac1-acfe-51ae7bb53345","shared_citers":8},{"title":"DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models","work_id":"c5006563-f3ec-438a-9e35-b7b484f34828","shared_citers":7},{"title":"LLaMA: Open and Efficient Foundation Language Models","work_id":"c018fc23-6f3f-4035-9d02-28a2173b2b9d","shared_citers":7},{"title":"Qwen3-VL Technical Report","work_id":"1fe243aa-e3c0-4da6-b391-4cbcfc88d5c0","shared_citers":7},{"title":"Qwen-Audio: Advancing Universal Audio Understanding via Unified Large-Scale Audio-Language Models","work_id":"d3f033ac-bfa8-4143-9d0d-51f3f5bd3f0e","shared_citers":7},{"title":"Qwen Technical Report","work_id":"bb1fd52f-6b2f-437c-9516-37bdf6eb9be8","shared_citers":7}],"time_series":[{"n":1,"year":2025},{"n":67,"year":2026}],"dependency_candidates":[]},"authors":[{"id":"d939bd6b-83ae-4395-8119-a82cb1101e6d","orcid":null,"display_name":"Hangrui Hu","source":"manual","import_confidence":0.72},{"id":"c42757d0-ace5-480f-b243-890854f9054c","orcid":null,"display_name":"Jin Xu","source":"manual","import_confidence":0.72},{"id":"c6a51477-94c9-40f9-bdd9-97f195765b54","orcid":null,"display_name":"Jinzheng He","source":"manual","import_confidence":0.72},{"id":"2bc8cfa9-c2ce-48ef-87d4-e0de52ae9a92","orcid":null,"display_name":"Shuai Bai","source":"manual","import_confidence":0.72},{"id":"64156499-4cb1-4de8-9683-9d46746db822","orcid":null,"display_name":"Ting He","source":"manual","import_confidence":0.72},{"id":"6fe3c884-415c-45f2-b8fb-6629cd6c24df","orcid":null,"display_name":"Zhifang Guo","source":"manual","import_confidence":0.72}]}}