保险文本视觉认知问答竞赛（Baseline）

本次赛题围绕保险扫描文档的OCR识别与智能问答展开，提供含票据等20多种类型的扫描文件数据集，含训练集5000余张图片及4万余个问答标注，测试集1000张左右图片及7000个问题。基线采用两阶段处理，先用PaddleOCR识别文本，再用PaddleNLP通过抽取式阅读理解得出答案，还给出了模型训练等相关内容及示例问答。

☞☞☞AI 智能聊天, 问答助手, AI 智能搜索, 免费无限量使用 DeepSeek R1 模型☜☜☜

赛题简介

在寿险、产险、健康险等保险的理赔流程和客户服务环节中，存在大量扫描文档，例如医疗票据、费用清单、病例等。对这些扫描文档进行文字检测与识别，并且提取出结构化信息，可以用于极速理赔、个人健康管理等业务场景。本次赛题将提供扫描文件数据集，参赛队利用OCR技术自动识别影像资料后，再通过AI智能判断所识别文字的内在逻辑，回答关于图片的自然语言问题。问题的答案是可以从图片中提取的任何文本/标记。

数据简介

本次大赛提供的数据集使用的扫描文件类型包括票据、说明、报告等20 多种。混合了印刷、打字和手写的内容。训练集有5000余张左右原始扫描文件及对应的 4万余个自然语言问答标注。提供的数据均已做了标注及脱敏。训练集数据包括：image：包含所有原始扫描文件图像train.csv：问答训练库，包含序号（index）、问题 ID(quesiton_id)、图片名称(filename)、问题（question_text)、答案(answer_text)，共 5 列readme：数据说明文档测试集数据规模为1000张左右原始扫描文件及对应的7000个自然语言问题，数据内容样例同训练集。测试集包含以下3个文件：image：包含所有原始扫描文件图像test1.csv：问答测试库，包含序号（index）、问题 ID(quesiton_id)、图片路径(filename)、问题（question_text)，共 4 列readme：数据说明文档

数据样例

样例一：

提问：西药费的金额是多少？ 回答：140.16 提问：140.16元购买了什么药品？ 回答：{甲}缘沙坦胶囊{基}

样例二：

无涯·问知

无涯·问知，是一款基于星环大模型底座，结合个人知识库、企业知识库、法律法规、财经等多种知识源的企业级垂直领域问答产品

40 查看详情

提问：这是一份关于什么药品的说明？ 回答：十三味疏肝胶囊 提问：药品的有效期是多久？ 回答：1.5年

基线总体架构

基线项目使用的是两阶段的处理方式：PaddleOCR：图像 ==OCR==> 文本信息PaddleNLP：文本信息 + 问题 ==抽取式阅读理解==> 答案

安装依赖包

PaddleOCR：GitHub、GiteePaddleNLP：GitHub、GiteeIn [ ]

!pip install paddleocr==2.0.4 paddlenlp==2.0.0rc18

解压官方数据集

如果需要自行处理数据，可以使用下方的命令进行解压In [ ]

# !tar -xf data/data83016/dataset.tar -C data

数据集预处理

使用 PaddleOCR 识别图片中的文本信息，将其作为抽取式阅读理解的文章并将数据格式转换为 SQuAD 格式，方便后续 PaddleNLP 读取使用处理完成的数据集以存放于 AIStudio 数据集中，可以直接挂载使用如果需要重新生成数据集的话，可以运行下面的命令，处理需耗时一个半小时左右，请耐心等待In [ ]

# !python gen_dataset.py

模型训练

基于 PaddleNLP 开发的抽取式阅读理解更多详细介绍请参考：『NLP经典项目集』08: 使用预训练模型完成阅读理解In [1]

!mkdir /home/aistudio/checkpointimport paddleimport paddlenlp as ppnlpfrom functools import partialfrom paddlenlp.data import Stack, Dict, Padfrom utils import prepare_train_features, prepare_validation_features, evaluate############参数配置################ 模型名称MODEL_NAME = "ernie-1.0"# 最大文本长度max_seq_length = 512# 文本滑动窗口步幅doc_stride = 128# 训练过程中的最大学习率learning_rate = 3e-5 # 训练轮次epochs = 1# 数据批次大小batch_size = 8# 学习率预热比例warmup_proportion = 0.1# 权重衰减系数，类似模型正则项策略，避免模型过拟合weight_decay = 0.01#############模型################# 加载模型model = ppnlp.transformers.ErnieForQuestionAnswering.from_pretrained(MODEL_NAME)# 加载 tokenizertokenizer = ppnlp.transformers.ErnieTokenizer.from_pretrained(MODEL_NAME)#############数据################ 加载数据集train_ds = ppnlp.datasets.load_dataset('dureader_robust', data_files='data/data83268/train.json')dev_ds = ppnlp.datasets.load_dataset('dureader_robust', data_files='data/data83268/dev.json')# 数据滑窗处理train_trans_func = partial(prepare_train_features,                            max_seq_length=max_seq_length,                            doc_stride=doc_stride,                           tokenizer=tokenizer)train_ds.map(train_trans_func, batched=True)dev_trans_func = partial(prepare_validation_features,                            max_seq_length=max_seq_length,                            doc_stride=doc_stride,                           tokenizer=tokenizer)                           dev_ds.map(dev_trans_func, batched=True)# 数据读取器配置train_batch_sampler = paddle.io.DistributedBatchSampler(        train_ds, batch_size=batch_size, shuffle=True)train_batchify_fn = lambda samples, fn=Dict({    "input_ids": Pad(axis=0, pad_val=tokenizer.pad_token_id),    "token_type_ids": Pad(axis=0, pad_val=tokenizer.pad_token_type_id),    "start_positions": Stack(dtype="int64"),    "end_positions": Stack(dtype="int64")}): fn(samples)train_data_loader = paddle.io.DataLoader(    dataset=train_ds,    batch_sampler=train_batch_sampler,    collate_fn=train_batchify_fn,    return_list=True)dev_batch_sampler = paddle.io.BatchSampler(    dev_ds, batch_size=batch_size, shuffle=False)dev_batchify_fn = lambda samples, fn=Dict({    "input_ids": Pad(axis=0, pad_val=tokenizer.pad_token_id),    "token_type_ids": Pad(axis=0, pad_val=tokenizer.pad_token_type_id)}): fn(samples)dev_data_loader = paddle.io.DataLoader(    dataset=dev_ds,    batch_sampler=dev_batch_sampler,    collate_fn=dev_batchify_fn,    return_list=True)#############优化器配置############## 学习率策略num_training_steps = len(train_data_loader) * epochslr_scheduler = ppnlp.transformers.LinearDecayWithWarmup(learning_rate, num_training_steps, warmup_proportion)# Generate parameter names needed to perform weight decay.# All bias and LayerNorm parameters are excluded.decay_params = [    p.name for n, p in model.named_parameters()    if not any(nd in n for nd in ["bias", "norm"])]# 设置优化器optimizer = paddle.optimizer.AdamW(    learning_rate=lr_scheduler,    parameters=model.parameters(),    weight_decay=weight_decay,    apply_decay_param_fun=lambda x: x in decay_params)#############损失函数################class CrossEntropyLossForSQuAD(paddle.nn.Layer):    def __init__(self):        super(CrossEntropyLossForSQuAD, self).__init__()    def forward(self, y, label):        start_logits, end_logits = y   # both shape are [batch_size, seq_len]        start_position, end_position = label        start_position = paddle.unsqueeze(start_position, axis=-1)        end_position = paddle.unsqueeze(end_position, axis=-1)        start_loss = paddle.nn.functional.softmax_with_cross_entropy(            logits=start_logits, label=start_position, soft_label=False)        start_loss = paddle.mean(start_loss)        end_loss = paddle.nn.functional.softmax_with_cross_entropy(            logits=end_logits, label=end_position, soft_label=False)        end_loss = paddle.mean(end_loss)        loss = (start_loss + end_loss) / 2        return loss#############模型训练################# 实例化 losscriterion = CrossEntropyLossForSQuAD()global_step = 0# 训练for epoch in range(1, epochs + 1):    for step, batch in enumerate(train_data_loader, start=1):        global_step += 1        input_ids, segment_ids, start_positions, end_positions = batch        logits = model(input_ids=input_ids, token_type_ids=segment_ids)        loss = criterion(logits, (start_positions, end_positions))        if global_step % 100 == 0 :            print("global step %d, epoch: %d, batch: %d, loss: %.5f" % (global_step, epoch, step, loss))        loss.backward()        optimizer.step()        lr_scheduler.step()        optimizer.clear_grad()    evaluate(model=model, data_loader=dev_data_loader) # 保存model.save_pretrained('/home/aistudio/checkpoint')tokenizer.save_pretrained('/home/aistudio/checkpoint')

[2021-04-22 20:48:38,873] [    INFO] - Already cached /home/aistudio/.paddlenlp/models/ernie-1.0/ernie_v1_chn_base.pdparams/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/layers.py:1303: UserWarning: Skip loading for classifier.weight. classifier.weight is not found in the provided dict.  warnings.warn(("Skip loading for {}. ".format(key) + str(err)))/opt/conda/envs/python35-paddle120-env/lib/python3.7/site-packages/paddle/fluid/dygraph/layers.py:1303: UserWarning: Skip loading for classifier.bias. classifier.bias is not found in the provided dict.  warnings.warn(("Skip loading for {}. ".format(key) + str(err)))[2021-04-22 20:48:42,972] [    INFO] - Found /home/aistudio/.paddlenlp/models/ernie-1.0/vocab.txt

global step 100, epoch: 1, batch: 100, loss: 5.33133global step 200, epoch: 1, batch: 200, loss: 2.81528global step 300, epoch: 1, batch: 300, loss: 1.96900global step 400, epoch: 1, batch: 400, loss: 1.99122global step 500, epoch: 1, batch: 500, loss: 2.26535global step 600, epoch: 1, batch: 600, loss: 1.91357global step 700, epoch: 1, batch: 700, loss: 1.60655global step 800, epoch: 1, batch: 800, loss: 1.76000global step 900, epoch: 1, batch: 900, loss: 1.13124global step 1000, epoch: 1, batch: 1000, loss: 1.72126global step 1100, epoch: 1, batch: 1100, loss: 1.89857global step 1200, epoch: 1, batch: 1200, loss: 1.47760global step 1300, epoch: 1, batch: 1300, loss: 1.70778global step 1400, epoch: 1, batch: 1400, loss: 1.30835global step 1500, epoch: 1, batch: 1500, loss: 2.72890global step 1600, epoch: 1, batch: 1600, loss: 1.84454global step 1700, epoch: 1, batch: 1700, loss: 3.09311global step 1800, epoch: 1, batch: 1800, loss: 1.83173global step 1900, epoch: 1, batch: 1900, loss: 1.07240global step 2000, epoch: 1, batch: 2000, loss: 1.33060global step 2100, epoch: 1, batch: 2100, loss: 1.04376global step 2200, epoch: 1, batch: 2200, loss: 1.63946global step 2300, epoch: 1, batch: 2300, loss: 2.03573global step 2400, epoch: 1, batch: 2400, loss: 1.47289global step 2500, epoch: 1, batch: 2500, loss: 1.08369global step 2600, epoch: 1, batch: 2600, loss: 1.38365global step 2700, epoch: 1, batch: 2700, loss: 1.71040global step 2800, epoch: 1, batch: 2800, loss: 1.26852global step 2900, epoch: 1, batch: 2900, loss: 2.52206global step 3000, epoch: 1, batch: 3000, loss: 1.91332global step 3100, epoch: 1, batch: 3100, loss: 1.47257global step 3200, epoch: 1, batch: 3200, loss: 1.06718global step 3300, epoch: 1, batch: 3300, loss: 1.79864global step 3400, epoch: 1, batch: 3400, loss: 1.58367global step 3500, epoch: 1, batch: 3500, loss: 0.83910global step 3600, epoch: 1, batch: 3600, loss: 1.63214global step 3700, epoch: 1, batch: 3700, loss: 3.27789global step 3800, epoch: 1, batch: 3800, loss: 1.13076global step 3900, epoch: 1, batch: 3900, loss: 1.18562global step 4000, epoch: 1, batch: 4000, loss: 0.91027global step 4100, epoch: 1, batch: 4100, loss: 0.81818global step 4200, epoch: 1, batch: 4200, loss: 1.16851global step 4300, epoch: 1, batch: 4300, loss: 1.64349global step 4400, epoch: 1, batch: 4400, loss: 1.51092global step 4500, epoch: 1, batch: 4500, loss: 2.32444global step 4600, epoch: 1, batch: 4600, loss: 1.04382global step 4700, epoch: 1, batch: 4700, loss: 1.18952global step 4800, epoch: 1, batch: 4800, loss: 1.08606global step 4900, epoch: 1, batch: 4900, loss: 1.37461global step 5000, epoch: 1, batch: 5000, loss: 1.14658global step 5100, epoch: 1, batch: 5100, loss: 1.24930global step 5200, epoch: 1, batch: 5200, loss: 0.97293global step 5300, epoch: 1, batch: 5300, loss: 1.39240global step 5400, epoch: 1, batch: 5400, loss: 1.52307global step 5500, epoch: 1, batch: 5500, loss: 1.01953global step 5600, epoch: 1, batch: 5600, loss: 1.54944global step 5700, epoch: 1, batch: 5700, loss: 1.86738global step 5800, epoch: 1, batch: 5800, loss: 1.54679global step 5900, epoch: 1, batch: 5900, loss: 2.57512global step 6000, epoch: 1, batch: 6000, loss: 1.68195global step 6100, epoch: 1, batch: 6100, loss: 2.33640global step 6200, epoch: 1, batch: 6200, loss: 1.33415global step 6300, epoch: 1, batch: 6300, loss: 1.53034global step 6400, epoch: 1, batch: 6400, loss: 2.18684global step 6500, epoch: 1, batch: 6500, loss: 1.03164global step 6600, epoch: 1, batch: 6600, loss: 1.31069global step 6700, epoch: 1, batch: 6700, loss: 1.56807global step 6800, epoch: 1, batch: 6800, loss: 0.98548global step 6900, epoch: 1, batch: 6900, loss: 0.99514global step 7000, epoch: 1, batch: 7000, loss: 0.98318global step 7100, epoch: 1, batch: 7100, loss: 1.00131global step 7200, epoch: 1, batch: 7200, loss: 0.95227global step 7300, epoch: 1, batch: 7300, loss: 1.08113global step 7400, epoch: 1, batch: 7400, loss: 0.82864global step 7500, epoch: 1, batch: 7500, loss: 2.03780global step 7600, epoch: 1, batch: 7600, loss: 1.08267global step 7700, epoch: 1, batch: 7700, loss: 1.19368global step 7800, epoch: 1, batch: 7800, loss: 1.13193global step 7900, epoch: 1, batch: 7900, loss: 0.86742global step 8000, epoch: 1, batch: 8000, loss: 1.33992global step 8100, epoch: 1, batch: 8100, loss: 2.19699global step 8200, epoch: 1, batch: 8200, loss: 0.98966global step 8300, epoch: 1, batch: 8300, loss: 0.91852global step 8400, epoch: 1, batch: 8400, loss: 0.98416global step 8500, epoch: 1, batch: 8500, loss: 0.93930global step 8600, epoch: 1, batch: 8600, loss: 1.14956global step 8700, epoch: 1, batch: 8700, loss: 0.98243global step 8800, epoch: 1, batch: 8800, loss: 1.07073global step 8900, epoch: 1, batch: 8900, loss: 0.87538global step 9000, epoch: 1, batch: 9000, loss: 1.29235global step 9100, epoch: 1, batch: 9100, loss: 1.42117global step 9200, epoch: 1, batch: 9200, loss: 2.06677global step 9300, epoch: 1, batch: 9300, loss: 1.20705global step 9400, epoch: 1, batch: 9400, loss: 1.14359global step 9500, epoch: 1, batch: 9500, loss: 0.92873global step 9600, epoch: 1, batch: 9600, loss: 1.21142global step 9700, epoch: 1, batch: 9700, loss: 1.35645global step 9800, epoch: 1, batch: 9800, loss: 1.16116global step 9900, epoch: 1, batch: 9900, loss: 1.08292global step 10000, epoch: 1, batch: 10000, loss: 1.59773global step 10100, epoch: 1, batch: 10100, loss: 1.01784global step 10200, epoch: 1, batch: 10200, loss: 0.67115global step 10300, epoch: 1, batch: 10300, loss: 1.47989global step 10400, epoch: 1, batch: 10400, loss: 1.01132global step 10500, epoch: 1, batch: 10500, loss: 0.97569global step 10600, epoch: 1, batch: 10600, loss: 1.14948global step 10700, epoch: 1, batch: 10700, loss: 2.03889global step 10800, epoch: 1, batch: 10800, loss: 1.08176global step 10900, epoch: 1, batch: 10900, loss: 0.78584global step 11000, epoch: 1, batch: 11000, loss: 2.09304global step 11100, epoch: 1, batch: 11100, loss: 2.07693global step 11200, epoch: 1, batch: 11200, loss: 1.08243global step 11300, epoch: 1, batch: 11300, loss: 1.74269global step 11400, epoch: 1, batch: 11400, loss: 2.41344global step 11500, epoch: 1, batch: 11500, loss: 0.73077global step 11600, epoch: 1, batch: 11600, loss: 0.81114global step 11700, epoch: 1, batch: 11700, loss: 1.29751global step 11800, epoch: 1, batch: 11800, loss: 1.33166global step 11900, epoch: 1, batch: 11900, loss: 0.89963global step 12000, epoch: 1, batch: 12000, loss: 0.94474global step 12100, epoch: 1, batch: 12100, loss: 1.06279global step 12200, epoch: 1, batch: 12200, loss: 1.91975global step 12300, epoch: 1, batch: 12300, loss: 1.00609global step 12400, epoch: 1, batch: 12400, loss: 1.47376global step 12500, epoch: 1, batch: 12500, loss: 1.03436global step 12600, epoch: 1, batch: 12600, loss: 1.01267global step 12700, epoch: 1, batch: 12700, loss: 1.22741global step 12800, epoch: 1, batch: 12800, loss: 1.01167global step 12900, epoch: 1, batch: 12900, loss: 2.15446global step 13000, epoch: 1, batch: 13000, loss: 0.77935global step 13100, epoch: 1, batch: 13100, loss: 1.25362global step 13200, epoch: 1, batch: 13200, loss: 1.98043global step 13300, epoch: 1, batch: 13300, loss: 1.87204global step 13400, epoch: 1, batch: 13400, loss: 1.13598global step 13500, epoch: 1, batch: 13500, loss: 1.03505global step 13600, epoch: 1, batch: 13600, loss: 0.94357global step 13700, epoch: 1, batch: 13700, loss: 0.98602global step 13800, epoch: 1, batch: 13800, loss: 0.88241global step 13900, epoch: 1, batch: 13900, loss: 1.53893global step 14000, epoch: 1, batch: 14000, loss: 1.36677global step 14100, epoch: 1, batch: 14100, loss: 1.08053global step 14200, epoch: 1, batch: 14200, loss: 1.37873global step 14300, epoch: 1, batch: 14300, loss: 0.66778global step 14400, epoch: 1, batch: 14400, loss: 2.18860global step 14500, epoch: 1, batch: 14500, loss: 1.57532global step 14600, epoch: 1, batch: 14600, loss: 0.99812global step 14700, epoch: 1, batch: 14700, loss: 0.86738global step 14800, epoch: 1, batch: 14800, loss: 1.23389global step 14900, epoch: 1, batch: 14900, loss: 1.15881global step 15000, epoch: 1, batch: 15000, loss: 1.03445global step 15100, epoch: 1, batch: 15100, loss: 0.88822global step 15200, epoch: 1, batch: 15200, loss: 1.13733global step 15300, epoch: 1, batch: 15300, loss: 1.28856global step 15400, epoch: 1, batch: 15400, loss: 1.17445global step 15500, epoch: 1, batch: 15500, loss: 1.28670global step 15600, epoch: 1, batch: 15600, loss: 2.49681global step 15700, epoch: 1, batch: 15700, loss: 1.19437global step 15800, epoch: 1, batch: 15800, loss: 1.06376global step 15900, epoch: 1, batch: 15900, loss: 0.98734global step 16000, epoch: 1, batch: 16000, loss: 1.17667global step 16100, epoch: 1, batch: 16100, loss: 1.28779global step 16200, epoch: 1, batch: 16200, loss: 1.05283global step 16300, epoch: 1, batch: 16300, loss: 1.62172global step 16400, epoch: 1, batch: 16400, loss: 0.92708global step 16500, epoch: 1, batch: 16500, loss: 0.95624global step 16600, epoch: 1, batch: 16600, loss: 1.29848global step 16700, epoch: 1, batch: 16700, loss: 1.27211global step 16800, epoch: 1, batch: 16800, loss: 1.17851global step 16900, epoch: 1, batch: 16900, loss: 1.28291global step 17000, epoch: 1, batch: 17000, loss: 1.08720global step 17100, epoch: 1, batch: 17100, loss: 1.08356global step 17200, epoch: 1, batch: 17200, loss: 1.00867Processing example: 1000time per 1000: 11.201786994934082Processing example: 2000time per 1000: 11.235816478729248Processing example: 3000time per 1000: 10.834845066070557Processing example: 4000time per 1000: 11.04150128364563Processing example: 5000time per 1000: 11.004519701004028Processing example: 6000time per 1000: 11.003149509429932Processing example: 7000time per 1000: 11.149619340896606{  "exact": 56.03663613655287,  "f1": 72.53400335174827,  "total": 1201,  "HasAns_exact": 56.03663613655287,  "HasAns_f1": 72.53400335174827,  "HasAns_total": 1201}问题： 本次医保范围支付多少钱？原文： 54020292北京市医疗网珍收费票据医保已世结發部监NO财16139-54-02实时结算：★医疗机构类型：交易流水号：2411000107180415993045社会保障卡号40096415918041502915城镇工男医保类型：单价数量单位业务流水号：性别：15380等级项目／规格姓名：金额有自作数量/单位鸡7500单价中成药贸6.2Y项目规格无自付：复方甲氧那明胶/48粒23.75001/瓶12.E200西药费收都联153.8000付jia酸左氧沙星/0.116.2G00无苏黄止咳囊/Q.45g2粒76.90002/津有效遣夫不北京市财政局印制·20172收费专用道172.32自付一17232000172.32起村金额17.750.G0衣饮医保范内金狮1332.51封顶金额0.00门诊大额支付0.0自付二0.累计医供内范金额190.07退体补充支付0.00年门诊大额票计支付0.白费个人支付金额陵军补财支付0.00190.070.09本饮支付后·个人账户余额单位补充险[原公疗]支付个人账户支付0.00基金支情2合计（大写收款人收款单位（章）答案： 172.32问题： 9260是什么的编号？原文： 54020292北京市医疗网珍收费票据医保已世结發部监NO财16139-54-02实时结算：★医疗机构类型：交易流水号：2411000107180415993045社会保障卡号40096415918041502915城镇工男医保类型：单价数量单位业务流水号：性别：15380等级项目／规格姓名：金额有自作数量/单位鸡7500单价中成药贸6.2Y项目规格无自付：复方甲氧那明胶/48粒23.75001/瓶12.E200西药费收都联153.8000付jia酸左氧沙星/0.116.2G00无苏黄止咳囊/Q.45g2粒76.90002/津有效遣夫不北京市财政局印制·20172收费专用道172.32自付一17232000172.32起村金额17.750.G0衣饮医保范内金狮1332.51封顶金额0.00门诊大额支付0.0自付二0.累计医供内范金额190.07退体补充支付0.00年门诊大额票计支付0.白费个人支付金额陵军补财支付0.00190.070.09本饮支付后·个人账户余额单位补充险[原公疗]支付个人账户支付0.00基金支情2合计（大写收款人收款单位（章）答案： 收款单位问题： 图7是表达什么的？原文： 东莞证券DONGGUANSECURITIES盛达资源（000603）深度报告45720062003.5520042002.53200222001.5112000.52002008-01-022013-11-022015-01-022015-08-022016-03-022016-10-022017-05-022017-12-022018-07-022019-02-022008-08-022010-12-022012-02-022012-09-022013-04-022014-06-022019-09-022009-10-022010-05-022011-07-022009-03-022020-04-022222SS30202Q90010S20C20美国：所有联储银行：资产：总资产美国：国债收益率：10年美国：联邦基金利率（日）美国：所有联储银行：资产：持有证券：美国国债资料来源：wind，东莞证券研究所资料来源：wind，东莞证券研究所图7：美国国债总额迅速增加（十亿美元）图8：美元流动性危机解除280005.0025.00260004.50240004.0020.00220003.503.002000015.002.50180002.001600010.00140001.501.00120005.00100000.5080000.002012-02-022013-11-022015-01-022016-03-022011-07-022013-04-022014-06-022015-08-022016-10-022017-05-022017-12-022008-01-022010-12-022012-09-022009-10-022010-05-022018-07-022008-08-022009-03-022019-02-022019-09-022012-09-022018-07-022008-01-022009-03-022009-10-022010-05-022010-12-022011-07-022012-02-022013-04-022013-11-022014-06-022015-08-022016-03-022017-05-022017-12-022019-02-022019-09-022020-04-022015-01-022016-10-022008-08-022020-04-02美国：国债总额-LIBOR：美元：3个月-美国：国债收益率：3个月M2同比增速（季调，右轴资料来源：wind，东莞证券研究所资料来源：wind，东莞证券研究所4.2.2美元处于下行通道，驱动黄金价格上行美国在疫情未受控情况下，强行重启经济，新冠新增感染人数仍在高位，市场对美国经济修复是否通畅存有忧虑。西欧及日本疫情率先于美国得到控制，美元相对欧元、英锈走弱。美国实施规模空前的财政刺激，导致政府负债迅速攀升，美元信用度下降美元指数下行将驱动以美元计价的黄金价格走高。图9：美国新冠新增感染人数仍在高位（人/日）图10：欧、日、美新冠新增感染人数（人/日）17请务必阅读末页声明。答案： 美国国债总额迅速增加（十亿美元）问题： 图8是说明什么的？原文： 东莞证券DONGGUANSECURITIES盛达资源（000603）深度报告45720062003.5520042002.53200222001.5112000.52002008-01-022013-11-022015-01-022015-08-022016-03-022016-10-022017-05-022017-12-022018-07-022019-02-022008-08-022010-12-022012-02-022012-09-022013-04-022014-06-022019-09-022009-10-022010-05-022011-07-022009-03-022020-04-022222SS30202Q90010S20C20美国：所有联储银行：资产：总资产美国：国债收益率：10年美国：联邦基金利率（日）美国：所有联储银行：资产：持有证券：美国国债资料来源：wind，东莞证券研究所资料来源：wind，东莞证券研究所图7：美国国债总额迅速增加（十亿美元）图8：美元流动性危机解除280005.0025.00260004.50240004.0020.00220003.503.002000015.002.50180002.001600010.00140001.501.00120005.00100000.5080000.002012-02-022013-11-022015-01-022016-03-022011-07-022013-04-022014-06-022015-08-022016-10-022017-05-022017-12-022008-01-022010-12-022012-09-022009-10-022010-05-022018-07-022008-08-022009-03-022019-02-022019-09-022012-09-022018-07-022008-01-022009-03-022009-10-022010-05-022010-12-022011-07-022012-02-022013-04-022013-11-022014-06-022015-08-022016-03-022017-05-022017-12-022019-02-022019-09-022020-04-022015-01-022016-10-022008-08-022020-04-02美国：国债总额-LIBOR：美元：3个月-美国：国债收益率：3个月M2同比增速（季调，右轴资料来源：wind，东莞证券研究所资料来源：wind，东莞证券研究所4.2.2美元处于下行通道，驱动黄金价格上行美国在疫情未受控情况下，强行重启经济，新冠新增感染人数仍在高位，市场对美国经济修复是否通畅存有忧虑。西欧及日本疫情率先于美国得到控制，美元相对欧元、英锈走弱。美国实施规模空前的财政刺激，导致政府负债迅速攀升，美元信用度下降美元指数下行将驱动以美元计价的黄金价格走高。图9：美国新冠新增感染人数仍在高位（人/日）图10：欧、日、美新冠新增感染人数（人/日）17请务必阅读末页声明。答案： 美元流动性危机解除问题： 文中上面的四张图的资料来源都是哪里？原文： 东莞证券DONGGUANSECURITIES盛达资源（000603）深度报告45720062003.5520042002.53200222001.5112000.52002008-01-022013-11-022015-01-022015-08-022016-03-022016-10-022017-05-022017-12-022018-07-022019-02-022008-08-022010-12-022012-02-022012-09-022013-04-022014-06-022019-09-022009-10-022010-05-022011-07-022009-03-022020-04-022222SS30202Q90010S20C20美国：所有联储银行：资产：总资产美国：国债收益率：10年美国：联邦基金利率（日）美国：所有联储银行：资产：持有证券：美国国债资料来源：wind，东莞证券研究所资料来源：wind，东莞证券研究所图7：美国国债总额迅速增加（十亿美元）图8：美元流动性危机解除280005.0025.00260004.50240004.0020.00220003.503.002000015.002.50180002.001600010.00140001.501.00120005.00100000.5080000.002012-02-022013-11-022015-01-022016-03-022011-07-022013-04-022014-06-022015-08-022016-10-022017-05-022017-12-022008-01-022010-12-022012-09-022009-10-022010-05-022018-07-022008-08-022009-03-022019-02-022019-09-022012-09-022018-07-022008-01-022009-03-022009-10-022010-05-022010-12-022011-07-022012-02-022013-04-022013-11-022014-06-022015-08-022016-03-022017-05-022017-12-022019-02-022019-09-022020-04-022015-01-022016-10-022008-08-022020-04-02美国：国债总额-LIBOR：美元：3个月-美国：国债收益率：3个月M2同比增速（季调，右轴资料来源：wind，东莞证券研究所资料来源：wind，东莞证券研究所4.2.2美元处于下行通道，驱动黄金价格上行美国在疫情未受控情况下，强行重启经济，新冠新增感染人数仍在高位，市场对美国经济修复是否通畅存有忧虑。西欧及日本疫情率先于美国得到控制，美元相对欧元、英锈走弱。美国实施规模空前的财政刺激，导致政府负债迅速攀升，美元信用度下降美元指数下行将驱动以美元计价的黄金价格走高。图9：美国新冠新增感染人数仍在高位（人/日）图10：欧、日、美新冠新增感染人数（人/日）17请务必阅读末页声明。答案： wind，东莞证券研究所

以上就是保险文本视觉认知问答竞赛（Baseline）的详细内容，更多请关注创想鸟其它相关文章！