Đọc hiểu Mức 3 – Đề 2

ĐỌC HIỂU · BÀI 1/4

Artificial intelligence in customs risk management

Read the passage and choose the best answer to each question.

Each day, customs administrations in major trading nations process tens of thousands of declarations. Physically inspecting every consignment would bring trade to a standstill, yet inspecting too few risks allowing contraband, undervalued goods and prohibited items to slip through. Customs agencies have long relied on risk profiles, sets of rules derived from experience and intelligence, to decide which shipments merit closer examination. Increasingly, these rule-based systems are being supplemented by machine learning.

Machine-learning models are trained on historical declarations and the outcomes of past inspections. By identifying subtle combinations of variables associated with violations, such as an unusual pairing of commodity code and declared value, or a new importer using a route favoured by known offenders, the models can assign each incoming declaration a probability of non-compliance. Several administrations report that such tools have substantially raised the "hit rate", the proportion of inspections that uncover irregularities.

Image analysis offers further possibilities. Algorithms can examine X-ray scans of containers and highlight areas where the image does not match the declared contents, assisting officers who must otherwise interpret hundreds of scans per shift.

Nonetheless, the technology raises legitimate concerns. Models trained on past inspection data may reproduce the biases embedded in earlier decisions: if certain types of traders were historically inspected more often, more violations will have been recorded for them, and the model may continue to target them disproportionately. Furthermore, many advanced models function as "black boxes", making it difficult to explain to a trader why a shipment was selected. Smugglers, too, adapt their behaviour once they suspect which patterns are being monitored.

For these reasons, most experts advocate a hybrid approach in which algorithms prioritise cases but experienced officers retain the final decision. Continuous monitoring of model performance, periodic retraining and transparency about general selection criteria are regarded as essential safeguards.

Câu 1/2000:00
Câu 1/20 · Bài 1

What ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠dilemma do customs administrations face, according to paragraph 1?

Câu 2/20 · Bài 1

What ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠is the "hit rate"?

Câu 3/20 · Bài 1

Why ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠might a model target some traders disproportionately?

Bạn đang làm thử 3/20 câu. Mua hoặc gia hạn gói để làm không giới hạn

17 câu còn lại dành cho tài khoản có gói
Bài 1 · Câu 4/20

What ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠does the term "black boxes" suggest about some models?

  1. They only work at night

  2. They are painted black

  3. Their decisions are difficult to explain

  4. They are very cheap

Bài 1 · Câu 5/20

What ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠approach do most experts advocate?

  1. Replacing officers entirely with algorithms

  2. Inspecting every container

  3. Stopping the use of machine learning

  4. Algorithms prioritising cases while officers make final decisions

Bài 2 · Câu 6/20

What ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠characterises the informal economy according to paragraph 1?

  1. It is found only in rural areas

  2. It consists of illegal activities

  3. It operates largely outside formal registration, contracts and social insurance

  4. It is run by large companies

Bài 2 · Câu 7/20

Which ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠is NOT mentioned as a function of the informal sector?

  1. Providing affordable goods to low-income households

  2. Funding public infrastructure

  3. Offering flexibility to workers

  4. Absorbing workers without formal jobs

Bài 2 · Câu 8/20

How ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠does a large informal sector affect the state?

  1. It narrows the tax base and limits social security coverage

  2. It widens the tax base

  3. It increases formal employment

  4. It reduces poverty automatically

Bài 2 · Câu 9/20

How ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠has policy thinking changed, according to paragraph 4?

  1. From cities to villages

  2. From lowering barriers to stricter enforcement

  3. From enforcement to lowering barriers to formalisation

  4. From taxation to subsidies

Bài 2 · Câu 10/20

The ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠word "tangible" in the last paragraph is closest in meaning to ______.

  1. temporary

  2. real and noticeable

  3. hidden

  4. theoretical

Bài 3 · Câu 11/20

Why ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠does the writer think traditional models are incomplete?

  1. Compliance is higher than they would predict

  2. People never evade tax

  3. Penalties are too high

  4. Audit rates are very high

Bài 3 · Câu 12/20

What ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠is a "nudge" according to the passage?

  1. A heavy penalty

  2. An audit

  3. A new tax law

  4. A low-cost intervention that steers behaviour without restricting choice

Bài 3 · Câu 13/20

Why ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠did the UK reminder letters work?

  1. They highlighted that most people had already paid on time

  2. They offered discounts

  3. They were sent by banks

  4. They threatened prison

Bài 3 · Câu 14/20

The ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠word "implicitly" in paragraph 2 is closest in meaning to ______.

  1. officially

  2. loudly

  3. without saying it directly

  4. incorrectly

Bài 3 · Câu 15/20

What ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠is the writer's overall attitude towards nudges?

  1. Strongly opposed

  2. Cautiously positive, seeing them as a useful complement

  3. Indifferent

  4. Believes they can replace enforcement

Bài 4 · Câu 16/20

What ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠causes the urban heat island effect?

  1. Fewer people in city centres

  2. Cold winds from the countryside

  3. More rain in cities

  4. Materials that absorb and store heat, plus waste heat and reduced ventilation

Bài 4 · Câu 17/20

What ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠"vicious cycle" does the writer describe?

  1. More trees cause more shade

  2. More air conditioning produces more waste heat, which increases demand for cooling

  3. More cars reduce temperatures

  4. More rain causes more floods

Bài 4 · Câu 18/20

Which ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠group is NOT mentioned as particularly at risk from heat?

  1. Office managers

  2. The elderly

  3. Outdoor workers

  4. Young children

Bài 4 · Câu 19/20

According ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠to the passage, when are mitigation measures most effective?

  1. When used only in summer

  2. When paid for by residents

  3. When added after buildings are completed

  4. When integrated into urban planning from the outset

Bài 4 · Câu 20/20

Why ⁠‍​‌‌‌​‌‌​​​​‌​​‌​​‌‌‌​‌‌​​​‌‌​​‌‌‌​​​‌‌‌‌‍⁠does the writer mention equity in the last paragraph?

  1. All neighbourhoods are equally affected

  2. Equity reduces temperatures

  3. Low-income neighbourhoods are often most exposed to extreme heat

  4. Rich areas have no trees