Hybrid challenge-response

1. A method for processing a message, comprising:

at an electronic computing device, receiving an electronic message via an electronic communications network;

classifying the message into one of a plurality of categories;

wherein the classification includes using content-based message analysis; and

challenging the message when the message is classified into a first category based at least in part on the content-based message analysis, wherein challenging the message includes:

determining a sender;

wherein the sender is associated with the message; and

sending a challenge message to the sender via the electronic communications network, wherein the challenge message includes a specification of a computational puzzle, wherein the computational puzzle includes a taxation function that can be solved to derive an answer to the computational puzzle, and wherein the challenge message includes information relating to a criterion for a satisfactory answer to the computational puzzle, wherein the information relating to the criterion for a satisfactory answer to the computational puzzle is formatted for automated interpretation by a computer.

2. The method of claim 1 , wherein the first category is associated with suspicious communications.

3. The method of claim 1 , wherein a second category is associated with legitimate communications.

4. The method of claim 3 , wherein a third category is associated with spam.

5. The method of claim 1 , wherein content-based message analysis includes using a text classifier.

6. The method of claim 5 , wherein the text classifier is a Bayesian text classifier.

7. The method of claim 1 , wherein classifying the message includes determining a likelihood that the message is associated with spam.

8. The method of claim 7 , wherein classifying the message includes comparing the likelihood to a threshold level.

9. The method of claim 8 , wherein the message is classified into a second category, wherein the second category is associated with legitimate communications, if it is determined that the likelihood does not exceed the threshold level.

10. The method of claim 8 , wherein the message is classified into the first category, if it is determined that the likelihood exceeds the threshold level.

11. The method of claim 8 , wherein classifying the message includes comparing the likelihood to a second threshold level, if the likelihood exceeds the threshold level.

12. The method of claim 11 , wherein classifying the message includes classifying the message into a third category, wherein the third category is associated with spam, if it is determined that the likelihood exceeds the second threshold level.

13. The method of claim 11 , wherein classifying the message includes classifying the message into the first category, if it is determined that the likelihood does not exceed the second threshold.

14. The method of claim 1 , wherein the message is an email message.

15. A system for processing a message, comprising:

a processor configured to:

receive an electronic message via an electronic communications network;

classify the message into one of a plurality of categories; wherein the classification includes using content-based message analysis; and

challenging the message when the message is classified into a first category based at least in part on the content-based message analysis, wherein challenging the message includes:

determining a sender;

wherein the sender is associated with the message; and

sending a challenge message to the sender via the electronic communications network, wherein the challenge message includes a specification of a computational puzzle, wherein the computational puzzle includes a taxation function that can be solved to derive an answer to the computational puzzle, and wherein the challenge message includes information relating to a criterion for a satisfactory answer to the computational puzzle, wherein the information relating to the criterion for a satisfactory answer to the computational puzzle is formatted for automated interpretation by a computer; and

a memory coupled with the processor, wherein the memory provides instructions to the processor.

16. A non-transitory computer readable storage medium, comprising a computer program product for processing a message, the computer program product comprising computer instructions for:

at an electronic computing device, receiving an electronic message via an electronic communications network;

classifying the message into one of a plurality of categories;

wherein the classification includes using content-based message analysis; and

challenging the message when the message is classified into a first category based at least on the content-based message analysis, wherein challenging the message includes:

determining a sender;

wherein the sender is associated with the message; and

sending a challenge message to the sender via the electronic communications network, wherein the challenge message includes a specification of a computational puzzle, wherein the computational puzzle includes a taxation function that can be solved to derive an answer to the computational puzzle, and wherein the challenge message includes information relating to a criterion for a satisfactory answer to the computational puzzle, wherein the information relating to the criterion for a satisfactory answer to the computational puzzle is formatted for automated interpretation by a computer.

17. A method for training a content classifier, comprising:

at an electronic computing device, receiving an electronic message via an electronic communications network;

issuing a challenge;

wherein issuing the challenge includes:

determining a sender;

wherein the sender is associated with the message; and

sending a challenge message to the sender via the electronic communications network, wherein the challenge message includes information relating to a criterion for a satisfactory response to the challenge message;

determining a result of the challenge; and

automatically reporting the result of the challenge to a content classifier.

18. The method of claim 17 , further comprising using the content classifier to filter spam from incoming messages.

19. The method of claim 17 , further comprising reporting to the content classifier whether the challenge message was delivered.

20. The method of claim 17 , further comprising reporting to the content classifier whether a response to the challenge message was received.

21. The method of claim 17 , further comprising reporting to the content classifier whether a correct response to the challenge message was received.

22. A system for training a content classifier, comprising:

a processor configured to:

receive an electronic message via an electronic communications network;

issue a challenge;

wherein issuing the challenge includes:

determining a sender;

wherein the sender is associated with the message; and

sending a challenge message to the sender via the electronic communications network, wherein the challenge message includes information relating to a criterion for a satisfactory response to the challenge message;

determine a result of the challenge; and

automatically report the result of the challenge to a content classifier; and

a memory coupled with the processor, wherein the memory provides instructions to the processor.

23. A non-transitory computer readable storage medium, comprising a computer program product for training a content classifier, the computer program product comprising computer instructions for:

receiving an electronic message via an electronic communications network;

issuing a challenge;

wherein issuing the challenge includes:

determining a sender;

wherein the sender is associated with the message; and

sending a challenge message to the sender via the electronic communications network, wherein the challenge message includes information relating to a criterion for a satisfactory response to the challenge message;

determining a result of the challenge; and

automatically reporting the result of the challenge to a content classifier.