Swarm_Intelligence_Flashcards

1
Q

What is the focus of the lecture on animal swarm intelligence?

A

Exploring collective behavior, swarm intelligence principles, and applications in swarm robotics and biological systems.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Three Basic Rules of Collective Behavior

A

1) Repulsion: Move away from very close neighbors.
2) Alignment: Match direction with nearby neighbors.
3) Attraction: Avoid isolation by staying close to others.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

How do simple local rules lead to complex group behavior?

A

Through repeated local interactions (repulsion, alignment, attraction), individuals self-organize and make coordinated group decisions.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Why is decentralization important in swarm intelligence?

A

No single leader or central processor; every individual has the same role, increasing adaptability, resilience, and reliability of the system.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

How does swarm intelligence differ from other group intelligence systems?

A

Swarm intelligence has no centralized leadership; all individuals share equal authority and information, unlike governments or matriarchal groups.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What happens if a key individual is removed from a decentralized swarm?

A

The system continues functioning effectively, because no single individual is critical for overall decision-making.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

Definition of Swarm Intelligence

A

Distributed, self-organized decision-making with information flowing through local interactions and no central control, enhancing group cognitive performance.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What are the four features of swarm intelligence?

A
  1. Adaptability, able to self organize
  2. Robustness, ablity to find new solution if current solutioin fails
  3. Reliability, individuals can be added or removed without disturbing the behaviour
  4. Simplicity, no central control
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

What is the “Many Wrongs Hypothesis” in group decision-making?

A
  • many wrongs hypothesis or the wisdom of the crowd
  • The idea that larger groups average out individual errors, leading to more accurate collective decisions (wisdom of the crowd).
  • The larger the group (given individuals contribute to decision making equally), the more likely to converge on the accurate decision
  • ocal interactions between individuals means on average, the outcome is more likely to be more correct (outcome is the average)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

How do bees use swarm intelligence to select a new hive location?

A

Each bee scouts potential sites, returns to waggle dance directions, and the swarm converges on the best average angle to move the queen.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Bee Waggle Dance Experiment

A

Demonstrates how many individuals each contribute local knowledge. The group aggregates these signals, producing a collective decision on where to move.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Why do larger groups often make more accurate decisions than individuals?

A

They combine multiple pieces of partial or error-prone information, averaging them out and reducing the impact of individual mistakes.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

In what scenario might an individual outperform a group?

A

When one individual has significantly superior knowledge or experience (e.g., an elephant matriarch), making a single informed decision potentially more accurate.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

What is the significance of “no individual contributes more to decision-making” in swarm systems?

A

It ensures resilience; removing any individual does not cripple the group’s overall function or decision accuracy.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Key Real-World Experiments

A
  • Bee waggle dance: Illustrates decentralized group choice.
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

What are some real-world applications of swarm intelligence?

A

Locust swarm studies, social insect behavior, swarm robotics, and human crowd-based problem solving.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

Why is it said that “the whole is greater than the sum of its parts” in swarm intelligence?

A

Individual contributions, though limited, collectively yield complex solutions that surpass what any single individual could achieve.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

How does swarm intelligence maintain reliability when members join or leave?

A

Distributed control means new or missing agents do not disrupt overall function, as decisions are made collectively without reliance on a central node.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

What key questions guide this lecture on group vs. individual decisions?

A

1) Do groups always outperform individuals?
2) How do groups arrive at better decisions?
3) Under what conditions can individuals outperform groups?

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
20
Q

What is the principle of “Fundamental Equality” in swarm systems?

A

In a true swarm, all individuals contribute equally to decision-making; no single agent should hold disproportionate influence or information.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
21
Q

what is a controversy of swarm intelligence recently found

A
  • found that some prior beliefs that bees show swarm intelligence may not be true
  • some study found that certain bees actually have more information than others, and is not actually swarm intelligence
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
22
Q

3D Camera Tracking of Bee Swarms

A

Researchers tracked thousands of bees using advanced camera arrays. Most flew at 2–3 m/s, but a minority flew at 9–10 m/s, acting as “streaker” bees that corrected the swarm’s speed and direction.

23
Q

Why do “streaker” bees challenge the idea of pure swarm intelligence?

A

They temporarily hold more information as they are flying faster and had been out to source the information, they influence other bees directions, hence not true swarm intelligence

24
Q

How do locust swarms transition from disorder to order?

A

Once enough locusts match velocity and direction, neighbors follow suit, shifting the group from disorganized to a cohesive, aligned swarm.

25
Q

What unique mechanism helps locusts maintain spacing in flight?

A

The threat of cannibalism. Locusts risk being bitten if they lag behind, encouraging them to keep moving in formation.

26
Q

Sensory Severance Experiment in Locusts

A

Researchers cut sensory nerves so locusts couldn’t feel bites. Without fear and sense of cannibalism, they failed to swarm, showing that cannibalism drives spacing and group movement.

27
Q

Why are ants considered a pervasive example of swarm intelligence?

A

Ants are highly social, prioritize group welfare, and exhibit decentralized decision-making through mechanisms like pheromone trails and quorum sensing.

28
Q

How do ants use pheromone trails to communicate?

A

They deposit chemical signals while walking. These trails guide others and get reinforced by repeated travel, forming a positive feedback loop.

29
Q

Why do ant colonies naturally find the shortest foraging routes?

A

Shorter paths allow ants to return more quickly and deposit stronger pheromone signals, which attracts more ants, reinforcing that route.

30
Q

Emergent Behavior in Ants

A

Complex solutions (like shortest path finding) arise from simple local interactions (pheromone deposition), without any central control or planning.

31
Q

What is the role of “quorum sensing” in ant nest site selection?

A

Ants only commit to moving the entire colony once enough individuals (a quorum or a threshold) have gathered and collectively approved a new site.

32
Q

How does tandem walking facilitate ant nest relocation?

A

A scout ant slowly leads another ant to a potential site via tandem walking. Once a critical number arrive (quorum), the colony shifts to rapidly carrying the rest of the nestmates, showing wisdom of the crowd (via quorum or threshold) converges to the best nest site

33
Q

how does quorum sensing and new nest site show swarm intelligence?

A

o Decentralized Information Integration:
* Not every ant has complete or identical information, but through repeated local interactions (the enticing of the tandem running), the collective aggregates these diverse inputs.
* The process ensures that the group’s decision reflects a consensus emerging from many partial observations, rather than relying on a single informed individual.
* Even if individual made error (i.e the nest site is not the best), the quorum sensing filters out noise (the ants won’t stay at the new nest site and is not enticed by the tandem walk), leading to more reliable outcomes

34
Q

How does the decoy effect reveal irrational decision-making?

A

Presenting a third, less attractive option can bias individual choices, making one of the original options suddenly seem more appealing, despite no real change in quality.

35
Q

Decoy Effect in Ant Colonies vs. Individuals

A

Individual ants can be tricked by the decoy effect due to cognitive overload. Colonies, however, use swarm intelligence to ignore the decoy and choose optimally.

36
Q

what is the experiment that shows ant colonies can ignore decoy effects?

A
  1. Ant colonies are offered a choice between two nest sites of equal quality:
    * one is darker but smaller entrance
    * one is lighter but bigger entrance
    * the two sites are of same quality, just trade-offs differently
    * they then added 2 decoy choices

they experimented with individual ant and then as a colony, to see if they make the irrational choice of picking the decoy

if the ants are not tricked by decoy they should choose all choices equally

in individual ants they show higher choice of the decoy, showing irrational choice

the reason for this is cognitive overload as they are presented with more choices above its cognitive capacity

in colonies they are immune to the decoy effect

this is because swarm intelligence of a colony dilutes individual cognitive overload, thereby negating the decoy effect and consistently favoring the optimal choice

37
Q

What is stigmergy?

A

An indirect coordination mechanism where agents leave traces (like pheromones) in the environment, influencing subsequent actions by other agents.

38
Q

How do ants illustrate stigmergy in foraging?

A

Each ant’s pheromone deposit updates the environment, guiding other ants to food sources and reinforcing efficient paths, creating coordinated foraging behaviours

39
Q

Why does swarm intelligence protect colonies from individual errors?

A

Decentralized interactions and quorum thresholds filter out mistakes, ensuring more robust and optimal group decisions.

40
Q

How do slime moulds demonstrate swarm intelligence in network formation?

A

Slime moulds, despite lacking a central brain, can optimize paths (e.g., replicating Tokyo’s subway network) through decentralized, adaptive behavior.

41
Q

Slime Mould Experimental Demonstration

A

Researchers placed slime mould in a dish with food sources positioned like Tokyo stations. The mould formed efficient connections mirroring actual subway routes, showing complex problem-solving without centralized control.

42
Q

When do groups typically make better decisions than individuals?

A

For complex or difficult tasks, collective input averages out errors (wisdom of the crowd), resulting in more accurate or effective decisions.

43
Q

How do groups integrate information to arrive at better decisions?

A

They rely on local interactions and feedback loops (positive or negative), allowing self-organization that combines multiple perspectives into a single outcome.

44
Q

Under what circumstances can an individual outperform a group?

A

When tasks are simpler or when each individual holds a different (and possibly superior) type of knowledge, a single informed individual may be more effective.

45
Q

Quorum Sensing (Recap)

A

A mechanism where a critical mass of individuals performing or endorsing a choice triggers a collective shift, enabling rapid and coordinated changes in group behavior.

46
Q

What is tandem walking in ants, and why is it important?

A

A scout ant carefully leads another to a new site. Once enough ants (a quorum) are convinced, they switch to rapidly transporting the rest, ensuring efficient nest relocation.

47
Q

How does the decoy effect illustrate cognitive overload in individual ants?

A

When offered multiple similar options, single ants can be swayed by an inferior “decoy.” Swarm intelligence, however, counters this bias by distributing decision-making.

48
Q

What role does environmental feedback play in swarm intelligence?

A

Agents leave or sense traces (e.g., pheromones) that shape the environment. Subsequent agents respond to these traces, reinforcing effective paths or behaviors.

49
Q

How do slime moulds reinforce the principle of decentralized optimization?

A

By showing that even a brainless organism can form efficient networks via local feedback, supporting the idea that complex solutions emerge from simple interactions.

50
Q

Conclusions on Group vs. Individual Decision-Making

A
  • Groups excel at difficult tasks via shared information.
  • Collective decisions form through self-organization and feedback.
  • Individuals may outperform groups on simpler tasks or when possessing unique expertise.
51
Q

What are the key features of swarm intelligence highlighted by slime moulds?

A

Decentralization, adaptability, and emergent optimization—no single controller, yet they form efficient pathways to resources.

52
Q

How do pheromone trails exemplify environmental feedback in ants?

A

Ants deposit pheromones as they travel. Stronger trails form on shorter or more successful routes, guiding others to follow and reinforcing efficiency.

53
Q

What is stigmergy and why is it crucial to swarm systems?

A

It’s indirect coordination via environmental cues (like pheromones). Agents leave traces that inform others, driving collective problem-solving without direct communication.

54
Q

Final Takeaways

A

Swarm intelligence emerges from local, decentralized interactions (bees, ants, locusts, slime moulds). While generally egalitarian, some agents may have more influence under certain conditions. Overall, groups often outperform individuals in complex tasks due to distributed information processing.