Study RoomSemester V · AY 2026–27

Past-paper question maps

See exactly what was asked.

Choose a subject to find each recovered question, its current syllabus unit, original marks and the linked answer.

58 distinct papers · 960 mapped question parts

These are inventory counts, not a claim that every recorded task is current or fully solvable. Historical exclusions, uncertain marks and a missing source page are shown below.

IS2

Intelligent Systems II

5 papers · 84 mapped parts

2024-04-17 to 2026-01-31

Original data errors and missing design choices are stated beside the answers. Conditional calculations use declared assumptions; no missing input is silently presented as a printed fact.

CN

Computer Networks

12 papers · 234 mapped parts

2022-11-16 to 2026-02-04

28 historical parts are outside the current syllabus; 1 part mixes current and historical content. One scan is missing page 6, leaving 1 recorded continuation unavailable. Same-date variants stay separate.

DL

Deep Learning

7 papers · 115 mapped parts

2022-11-19 to 2026-02-10

Two original parts mix current topics with historical tasks. Their original whole-part marks remain visible; the current-only share is not guessed. Missing numerical inputs and conventions are disclosed beside the solutions.

NLP

Natural Language Processing

8 papers · 133 mapped parts

2022-11-22 to 2026-02-09

8 individual allocations are unprinted. December 2023 offers 120 marks rather than 140. Doc2Vec is labelled as a historical named-method extension; numerical counting conventions remain explicit.

DM

Disaster Management

12 papers · 194 mapped parts

2022-12-01 to 2026-07-10

Every recorded part has printed marks. No standalone supplied-data numerical was found in these papers. A missing class-specific case narrative and unverified historical operation details remain qualified.

APR

Advertising & Public Relations

14 papers · 200 mapped parts

2022-12-02 to 2026-07-11

14 individual mark allocations are unprinted or conflicting. Faculty case narratives not supplied are identified in the answers. Media calculation examples inside explanations are distinguished from standalone PYQ numericals.

Read the marks correctly

Read the study content without JavaScript · AI reading index